diff --git a/vendor/github.com/go-redis/redis/v8/.gitignore b/vendor/github.com/go-redis/redis/v8/.gitignore
new file mode 100644
index 0000000..b975a7b
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/.gitignore
@@ -0,0 +1,3 @@
+*.rdb
+testdata/*/
+.idea/
diff --git a/vendor/github.com/go-redis/redis/v8/.golangci.yml b/vendor/github.com/go-redis/redis/v8/.golangci.yml
new file mode 100644
index 0000000..1e8d238
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/.golangci.yml
@@ -0,0 +1,21 @@
+run:
+  concurrency: 8
+  deadline: 5m
+  tests: false
+linters:
+  enable-all: true
+  disable:
+    - funlen
+    - gochecknoglobals
+    - gochecknoinits
+    - gocognit
+    - goconst
+    - godox
+    - gosec
+    - maligned
+    - wsl
+    - gomnd
+    - goerr113
+    - exhaustive
+    - nestif
+    - nlreturn
diff --git a/vendor/github.com/go-redis/redis/v8/.prettierrc b/vendor/github.com/go-redis/redis/v8/.prettierrc
new file mode 100644
index 0000000..8b7f044
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/.prettierrc
@@ -0,0 +1,4 @@
+semi: false
+singleQuote: true
+proseWrap: always
+printWidth: 100
diff --git a/vendor/github.com/go-redis/redis/v8/.travis.yml b/vendor/github.com/go-redis/redis/v8/.travis.yml
new file mode 100644
index 0000000..1bf578d
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/.travis.yml
@@ -0,0 +1,20 @@
+dist: xenial
+language: go
+
+services:
+  - redis-server
+
+go:
+  - 1.14.x
+  - 1.15.x
+  - tip
+
+matrix:
+  allow_failures:
+    - go: tip
+
+go_import_path: github.com/go-redis/redis
+
+before_install:
+  - curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s --
+    -b $(go env GOPATH)/bin v1.31.0
diff --git a/vendor/github.com/go-redis/redis/v8/CHANGELOG.md b/vendor/github.com/go-redis/redis/v8/CHANGELOG.md
new file mode 100644
index 0000000..8392d54
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/CHANGELOG.md
@@ -0,0 +1,5 @@
+# Changelog
+
+> :heart: [**Uptrace.dev** - distributed traces, logs, and errors in one place](https://uptrace.dev)
+
+See https://redis.uptrace.dev/changelog/
diff --git a/vendor/github.com/go-redis/redis/v8/LICENSE b/vendor/github.com/go-redis/redis/v8/LICENSE
new file mode 100644
index 0000000..298bed9
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/LICENSE
@@ -0,0 +1,25 @@
+Copyright (c) 2013 The github.com/go-redis/redis Authors.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+   * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+   * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/vendor/github.com/go-redis/redis/v8/Makefile b/vendor/github.com/go-redis/redis/v8/Makefile
new file mode 100644
index 0000000..49e4c96
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/Makefile
@@ -0,0 +1,21 @@
+all: testdeps
+	go test ./...
+	go test ./... -short -race
+	go test ./... -run=NONE -bench=. -benchmem
+	env GOOS=linux GOARCH=386 go test ./...
+	go vet
+	golangci-lint run
+
+testdeps: testdata/redis/src/redis-server
+
+bench: testdeps
+	go test ./... -test.run=NONE -test.bench=. -test.benchmem
+
+.PHONY: all test testdeps bench
+
+testdata/redis:
+	mkdir -p $@
+	wget -qO- http://download.redis.io/redis-stable.tar.gz | tar xvz --strip-components=1 -C $@
+
+testdata/redis/src/redis-server: testdata/redis
+	cd $< && make all
diff --git a/vendor/github.com/go-redis/redis/v8/README.md b/vendor/github.com/go-redis/redis/v8/README.md
new file mode 100644
index 0000000..da5d0fb
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/README.md
@@ -0,0 +1,137 @@
+# Redis client for Golang
+
+[![Build Status](https://travis-ci.org/go-redis/redis.png?branch=master)](https://travis-ci.org/go-redis/redis)
+[![PkgGoDev](https://pkg.go.dev/badge/github.com/go-redis/redis/v8)](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc)
+[![Documentation](https://img.shields.io/badge/redis-documentation-informational)](https://redis.uptrace.dev/)
+[![Chat](https://discordapp.com/api/guilds/752070105847955518/widget.png)](https://discord.gg/rWtp5Aj)
+
+> :heart: [**Uptrace.dev** - distributed traces, logs, and errors in one place](https://uptrace.dev)
+
+- Join [Discord](https://discord.gg/rWtp5Aj) to ask questions.
+- [Documentation](https://redis.uptrace.dev)
+- [Reference](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc)
+- [Examples](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#pkg-examples)
+- [RealWorld example app](https://github.com/uptrace/go-treemux-realworld-example-app)
+
+## Ecosystem
+
+- [Distributed Locks](https://github.com/bsm/redislock).
+- [Redis Cache](https://github.com/go-redis/cache).
+- [Rate limiting](https://github.com/go-redis/redis_rate).
+
+## Features
+
+- Redis 3 commands except QUIT, MONITOR, and SYNC.
+- Automatic connection pooling with
+  [circuit breaker](https://en.wikipedia.org/wiki/Circuit_breaker_design_pattern) support.
+- [Pub/Sub](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#PubSub).
+- [Transactions](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#example-Client-TxPipeline).
+- [Pipeline](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#example-Client-Pipeline) and
+  [TxPipeline](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#example-Client-TxPipeline).
+- [Scripting](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#Script).
+- [Timeouts](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#Options).
+- [Redis Sentinel](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#NewFailoverClient).
+- [Redis Cluster](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#NewClusterClient).
+- [Cluster of Redis Servers](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#example-NewClusterClient--ManualSetup)
+  without using cluster mode and Redis Sentinel.
+- [Ring](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#NewRing).
+- [Instrumentation](https://pkg.go.dev/github.com/go-redis/redis/v8?tab=doc#ex-package--Instrumentation).
+
+## Installation
+
+go-redis supports 2 last Go versions and requires a Go version with
+[modules](https://github.com/golang/go/wiki/Modules) support. So make sure to initialize a Go
+module:
+
+```shell
+go mod init github.com/my/repo
+```
+
+And then install go-redis (note _v8_ in the import; omitting it is a popular mistake):
+
+```shell
+go get github.com/go-redis/redis/v8
+```
+
+## Quickstart
+
+```go
+import (
+    "context"
+    "github.com/go-redis/redis/v8"
+)
+
+var ctx = context.Background()
+
+func ExampleClient() {
+    rdb := redis.NewClient(&redis.Options{
+        Addr:     "localhost:6379",
+        Password: "", // no password set
+        DB:       0,  // use default DB
+    })
+
+    err := rdb.Set(ctx, "key", "value", 0).Err()
+    if err != nil {
+        panic(err)
+    }
+
+    val, err := rdb.Get(ctx, "key").Result()
+    if err != nil {
+        panic(err)
+    }
+    fmt.Println("key", val)
+
+    val2, err := rdb.Get(ctx, "key2").Result()
+    if err == redis.Nil {
+        fmt.Println("key2 does not exist")
+    } else if err != nil {
+        panic(err)
+    } else {
+        fmt.Println("key2", val2)
+    }
+    // Output: key value
+    // key2 does not exist
+}
+```
+
+## Look and feel
+
+Some corner cases:
+
+```go
+// SET key value EX 10 NX
+set, err := rdb.SetNX(ctx, "key", "value", 10*time.Second).Result()
+
+// SET key value keepttl NX
+set, err := rdb.SetNX(ctx, "key", "value", redis.KeepTTL).Result()
+
+// SORT list LIMIT 0 2 ASC
+vals, err := rdb.Sort(ctx, "list", &redis.Sort{Offset: 0, Count: 2, Order: "ASC"}).Result()
+
+// ZRANGEBYSCORE zset -inf +inf WITHSCORES LIMIT 0 2
+vals, err := rdb.ZRangeByScoreWithScores(ctx, "zset", &redis.ZRangeBy{
+    Min: "-inf",
+    Max: "+inf",
+    Offset: 0,
+    Count: 2,
+}).Result()
+
+// ZINTERSTORE out 2 zset1 zset2 WEIGHTS 2 3 AGGREGATE SUM
+vals, err := rdb.ZInterStore(ctx, "out", &redis.ZStore{
+    Keys: []string{"zset1", "zset2"},
+    Weights: []int64{2, 3}
+}).Result()
+
+// EVAL "return {KEYS[1],ARGV[1]}" 1 "key" "hello"
+vals, err := rdb.Eval(ctx, "return {KEYS[1],ARGV[1]}", []string{"key"}, "hello").Result()
+
+// custom command
+res, err := rdb.Do(ctx, "set", "key", "value").Result()
+```
+
+## See also
+
+- [Fast and flexible HTTP router](https://github.com/vmihailenco/treemux)
+- [Golang PostgreSQL ORM](https://github.com/go-pg/pg)
+- [Golang msgpack](https://github.com/vmihailenco/msgpack)
+- [Golang message task queue](https://github.com/vmihailenco/taskq)
diff --git a/vendor/github.com/go-redis/redis/v8/cluster.go b/vendor/github.com/go-redis/redis/v8/cluster.go
new file mode 100644
index 0000000..a6ce5c5
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/cluster.go
@@ -0,0 +1,1697 @@
+package redis
+
+import (
+	"context"
+	"crypto/tls"
+	"fmt"
+	"math"
+	"net"
+	"runtime"
+	"sort"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/hashtag"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/proto"
+	"github.com/go-redis/redis/v8/internal/rand"
+)
+
+var errClusterNoNodes = fmt.Errorf("redis: cluster has no nodes")
+
+// ClusterOptions are used to configure a cluster client and should be
+// passed to NewClusterClient.
+type ClusterOptions struct {
+	// A seed list of host:port addresses of cluster nodes.
+	Addrs []string
+
+	// NewClient creates a cluster node client with provided name and options.
+	NewClient func(opt *Options) *Client
+
+	// The maximum number of retries before giving up. Command is retried
+	// on network errors and MOVED/ASK redirects.
+	// Default is 3 retries.
+	MaxRedirects int
+
+	// Enables read-only commands on slave nodes.
+	ReadOnly bool
+	// Allows routing read-only commands to the closest master or slave node.
+	// It automatically enables ReadOnly.
+	RouteByLatency bool
+	// Allows routing read-only commands to the random master or slave node.
+	// It automatically enables ReadOnly.
+	RouteRandomly bool
+
+	// Optional function that returns cluster slots information.
+	// It is useful to manually create cluster of standalone Redis servers
+	// and load-balance read/write operations between master and slaves.
+	// It can use service like ZooKeeper to maintain configuration information
+	// and Cluster.ReloadState to manually trigger state reloading.
+	ClusterSlots func(context.Context) ([]ClusterSlot, error)
+
+	// Following options are copied from Options struct.
+
+	Dialer func(ctx context.Context, network, addr string) (net.Conn, error)
+
+	OnConnect func(ctx context.Context, cn *Conn) error
+
+	Username string
+	Password string
+
+	MaxRetries      int
+	MinRetryBackoff time.Duration
+	MaxRetryBackoff time.Duration
+
+	DialTimeout  time.Duration
+	ReadTimeout  time.Duration
+	WriteTimeout time.Duration
+
+	// PoolSize applies per cluster node and not for the whole cluster.
+	PoolSize           int
+	MinIdleConns       int
+	MaxConnAge         time.Duration
+	PoolTimeout        time.Duration
+	IdleTimeout        time.Duration
+	IdleCheckFrequency time.Duration
+
+	TLSConfig *tls.Config
+}
+
+func (opt *ClusterOptions) init() {
+	if opt.MaxRedirects == -1 {
+		opt.MaxRedirects = 0
+	} else if opt.MaxRedirects == 0 {
+		opt.MaxRedirects = 3
+	}
+
+	if (opt.RouteByLatency || opt.RouteRandomly) && opt.ClusterSlots == nil {
+		opt.ReadOnly = true
+	}
+
+	if opt.PoolSize == 0 {
+		opt.PoolSize = 5 * runtime.NumCPU()
+	}
+
+	switch opt.ReadTimeout {
+	case -1:
+		opt.ReadTimeout = 0
+	case 0:
+		opt.ReadTimeout = 3 * time.Second
+	}
+	switch opt.WriteTimeout {
+	case -1:
+		opt.WriteTimeout = 0
+	case 0:
+		opt.WriteTimeout = opt.ReadTimeout
+	}
+
+	if opt.MaxRetries == 0 {
+		opt.MaxRetries = -1
+	}
+	switch opt.MinRetryBackoff {
+	case -1:
+		opt.MinRetryBackoff = 0
+	case 0:
+		opt.MinRetryBackoff = 8 * time.Millisecond
+	}
+	switch opt.MaxRetryBackoff {
+	case -1:
+		opt.MaxRetryBackoff = 0
+	case 0:
+		opt.MaxRetryBackoff = 512 * time.Millisecond
+	}
+
+	if opt.NewClient == nil {
+		opt.NewClient = NewClient
+	}
+}
+
+func (opt *ClusterOptions) clientOptions() *Options {
+	const disableIdleCheck = -1
+
+	return &Options{
+		Dialer:    opt.Dialer,
+		OnConnect: opt.OnConnect,
+
+		Username: opt.Username,
+		Password: opt.Password,
+
+		MaxRetries:      opt.MaxRetries,
+		MinRetryBackoff: opt.MinRetryBackoff,
+		MaxRetryBackoff: opt.MaxRetryBackoff,
+
+		DialTimeout:  opt.DialTimeout,
+		ReadTimeout:  opt.ReadTimeout,
+		WriteTimeout: opt.WriteTimeout,
+
+		PoolSize:           opt.PoolSize,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: disableIdleCheck,
+
+		readOnly: opt.ReadOnly,
+
+		TLSConfig: opt.TLSConfig,
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type clusterNode struct {
+	Client *Client
+
+	latency    uint32 // atomic
+	generation uint32 // atomic
+	failing    uint32 // atomic
+}
+
+func newClusterNode(clOpt *ClusterOptions, addr string) *clusterNode {
+	opt := clOpt.clientOptions()
+	opt.Addr = addr
+	node := clusterNode{
+		Client: clOpt.NewClient(opt),
+	}
+
+	node.latency = math.MaxUint32
+	if clOpt.RouteByLatency {
+		go node.updateLatency()
+	}
+
+	return &node
+}
+
+func (n *clusterNode) String() string {
+	return n.Client.String()
+}
+
+func (n *clusterNode) Close() error {
+	return n.Client.Close()
+}
+
+func (n *clusterNode) updateLatency() {
+	const numProbe = 10
+	var dur uint64
+
+	for i := 0; i < numProbe; i++ {
+		time.Sleep(time.Duration(10+rand.Intn(10)) * time.Millisecond)
+
+		start := time.Now()
+		n.Client.Ping(context.TODO())
+		dur += uint64(time.Since(start) / time.Microsecond)
+	}
+
+	latency := float64(dur) / float64(numProbe)
+	atomic.StoreUint32(&n.latency, uint32(latency+0.5))
+}
+
+func (n *clusterNode) Latency() time.Duration {
+	latency := atomic.LoadUint32(&n.latency)
+	return time.Duration(latency) * time.Microsecond
+}
+
+func (n *clusterNode) MarkAsFailing() {
+	atomic.StoreUint32(&n.failing, uint32(time.Now().Unix()))
+}
+
+func (n *clusterNode) Failing() bool {
+	const timeout = 15 // 15 seconds
+
+	failing := atomic.LoadUint32(&n.failing)
+	if failing == 0 {
+		return false
+	}
+	if time.Now().Unix()-int64(failing) < timeout {
+		return true
+	}
+	atomic.StoreUint32(&n.failing, 0)
+	return false
+}
+
+func (n *clusterNode) Generation() uint32 {
+	return atomic.LoadUint32(&n.generation)
+}
+
+func (n *clusterNode) SetGeneration(gen uint32) {
+	for {
+		v := atomic.LoadUint32(&n.generation)
+		if gen < v || atomic.CompareAndSwapUint32(&n.generation, v, gen) {
+			break
+		}
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type clusterNodes struct {
+	opt *ClusterOptions
+
+	mu          sync.RWMutex
+	addrs       []string
+	nodes       map[string]*clusterNode
+	activeAddrs []string
+	closed      bool
+
+	_generation uint32 // atomic
+}
+
+func newClusterNodes(opt *ClusterOptions) *clusterNodes {
+	return &clusterNodes{
+		opt: opt,
+
+		addrs: opt.Addrs,
+		nodes: make(map[string]*clusterNode),
+	}
+}
+
+func (c *clusterNodes) Close() error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.closed {
+		return nil
+	}
+	c.closed = true
+
+	var firstErr error
+	for _, node := range c.nodes {
+		if err := node.Client.Close(); err != nil && firstErr == nil {
+			firstErr = err
+		}
+	}
+
+	c.nodes = nil
+	c.activeAddrs = nil
+
+	return firstErr
+}
+
+func (c *clusterNodes) Addrs() ([]string, error) {
+	var addrs []string
+	c.mu.RLock()
+	closed := c.closed
+	if !closed {
+		if len(c.activeAddrs) > 0 {
+			addrs = c.activeAddrs
+		} else {
+			addrs = c.addrs
+		}
+	}
+	c.mu.RUnlock()
+
+	if closed {
+		return nil, pool.ErrClosed
+	}
+	if len(addrs) == 0 {
+		return nil, errClusterNoNodes
+	}
+	return addrs, nil
+}
+
+func (c *clusterNodes) NextGeneration() uint32 {
+	return atomic.AddUint32(&c._generation, 1)
+}
+
+// GC removes unused nodes.
+func (c *clusterNodes) GC(generation uint32) {
+	//nolint:prealloc
+	var collected []*clusterNode
+
+	c.mu.Lock()
+
+	c.activeAddrs = c.activeAddrs[:0]
+	for addr, node := range c.nodes {
+		if node.Generation() >= generation {
+			c.activeAddrs = append(c.activeAddrs, addr)
+			if c.opt.RouteByLatency {
+				go node.updateLatency()
+			}
+			continue
+		}
+
+		delete(c.nodes, addr)
+		collected = append(collected, node)
+	}
+
+	c.mu.Unlock()
+
+	for _, node := range collected {
+		_ = node.Client.Close()
+	}
+}
+
+func (c *clusterNodes) Get(addr string) (*clusterNode, error) {
+	node, err := c.get(addr)
+	if err != nil {
+		return nil, err
+	}
+	if node != nil {
+		return node, nil
+	}
+
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.closed {
+		return nil, pool.ErrClosed
+	}
+
+	node, ok := c.nodes[addr]
+	if ok {
+		return node, nil
+	}
+
+	node = newClusterNode(c.opt, addr)
+
+	c.addrs = appendIfNotExists(c.addrs, addr)
+	c.nodes[addr] = node
+
+	return node, nil
+}
+
+func (c *clusterNodes) get(addr string) (*clusterNode, error) {
+	var node *clusterNode
+	var err error
+	c.mu.RLock()
+	if c.closed {
+		err = pool.ErrClosed
+	} else {
+		node = c.nodes[addr]
+	}
+	c.mu.RUnlock()
+	return node, err
+}
+
+func (c *clusterNodes) All() ([]*clusterNode, error) {
+	c.mu.RLock()
+	defer c.mu.RUnlock()
+
+	if c.closed {
+		return nil, pool.ErrClosed
+	}
+
+	cp := make([]*clusterNode, 0, len(c.nodes))
+	for _, node := range c.nodes {
+		cp = append(cp, node)
+	}
+	return cp, nil
+}
+
+func (c *clusterNodes) Random() (*clusterNode, error) {
+	addrs, err := c.Addrs()
+	if err != nil {
+		return nil, err
+	}
+
+	n := rand.Intn(len(addrs))
+	return c.Get(addrs[n])
+}
+
+//------------------------------------------------------------------------------
+
+type clusterSlot struct {
+	start, end int
+	nodes      []*clusterNode
+}
+
+type clusterSlotSlice []*clusterSlot
+
+func (p clusterSlotSlice) Len() int {
+	return len(p)
+}
+
+func (p clusterSlotSlice) Less(i, j int) bool {
+	return p[i].start < p[j].start
+}
+
+func (p clusterSlotSlice) Swap(i, j int) {
+	p[i], p[j] = p[j], p[i]
+}
+
+type clusterState struct {
+	nodes   *clusterNodes
+	Masters []*clusterNode
+	Slaves  []*clusterNode
+
+	slots []*clusterSlot
+
+	generation uint32
+	createdAt  time.Time
+}
+
+func newClusterState(
+	nodes *clusterNodes, slots []ClusterSlot, origin string,
+) (*clusterState, error) {
+	c := clusterState{
+		nodes: nodes,
+
+		slots: make([]*clusterSlot, 0, len(slots)),
+
+		generation: nodes.NextGeneration(),
+		createdAt:  time.Now(),
+	}
+
+	originHost, _, _ := net.SplitHostPort(origin)
+	isLoopbackOrigin := isLoopback(originHost)
+
+	for _, slot := range slots {
+		var nodes []*clusterNode
+		for i, slotNode := range slot.Nodes {
+			addr := slotNode.Addr
+			if !isLoopbackOrigin {
+				addr = replaceLoopbackHost(addr, originHost)
+			}
+
+			node, err := c.nodes.Get(addr)
+			if err != nil {
+				return nil, err
+			}
+
+			node.SetGeneration(c.generation)
+			nodes = append(nodes, node)
+
+			if i == 0 {
+				c.Masters = appendUniqueNode(c.Masters, node)
+			} else {
+				c.Slaves = appendUniqueNode(c.Slaves, node)
+			}
+		}
+
+		c.slots = append(c.slots, &clusterSlot{
+			start: slot.Start,
+			end:   slot.End,
+			nodes: nodes,
+		})
+	}
+
+	sort.Sort(clusterSlotSlice(c.slots))
+
+	time.AfterFunc(time.Minute, func() {
+		nodes.GC(c.generation)
+	})
+
+	return &c, nil
+}
+
+func replaceLoopbackHost(nodeAddr, originHost string) string {
+	nodeHost, nodePort, err := net.SplitHostPort(nodeAddr)
+	if err != nil {
+		return nodeAddr
+	}
+
+	nodeIP := net.ParseIP(nodeHost)
+	if nodeIP == nil {
+		return nodeAddr
+	}
+
+	if !nodeIP.IsLoopback() {
+		return nodeAddr
+	}
+
+	// Use origin host which is not loopback and node port.
+	return net.JoinHostPort(originHost, nodePort)
+}
+
+func isLoopback(host string) bool {
+	ip := net.ParseIP(host)
+	if ip == nil {
+		return true
+	}
+	return ip.IsLoopback()
+}
+
+func (c *clusterState) slotMasterNode(slot int) (*clusterNode, error) {
+	nodes := c.slotNodes(slot)
+	if len(nodes) > 0 {
+		return nodes[0], nil
+	}
+	return c.nodes.Random()
+}
+
+func (c *clusterState) slotSlaveNode(slot int) (*clusterNode, error) {
+	nodes := c.slotNodes(slot)
+	switch len(nodes) {
+	case 0:
+		return c.nodes.Random()
+	case 1:
+		return nodes[0], nil
+	case 2:
+		if slave := nodes[1]; !slave.Failing() {
+			return slave, nil
+		}
+		return nodes[0], nil
+	default:
+		var slave *clusterNode
+		for i := 0; i < 10; i++ {
+			n := rand.Intn(len(nodes)-1) + 1
+			slave = nodes[n]
+			if !slave.Failing() {
+				return slave, nil
+			}
+		}
+
+		// All slaves are loading - use master.
+		return nodes[0], nil
+	}
+}
+
+func (c *clusterState) slotClosestNode(slot int) (*clusterNode, error) {
+	nodes := c.slotNodes(slot)
+	if len(nodes) == 0 {
+		return c.nodes.Random()
+	}
+
+	var node *clusterNode
+	for _, n := range nodes {
+		if n.Failing() {
+			continue
+		}
+		if node == nil || n.Latency() < node.Latency() {
+			node = n
+		}
+	}
+	if node != nil {
+		return node, nil
+	}
+
+	// If all nodes are failing - return random node
+	return c.nodes.Random()
+}
+
+func (c *clusterState) slotRandomNode(slot int) (*clusterNode, error) {
+	nodes := c.slotNodes(slot)
+	if len(nodes) == 0 {
+		return c.nodes.Random()
+	}
+	n := rand.Intn(len(nodes))
+	return nodes[n], nil
+}
+
+func (c *clusterState) slotNodes(slot int) []*clusterNode {
+	i := sort.Search(len(c.slots), func(i int) bool {
+		return c.slots[i].end >= slot
+	})
+	if i >= len(c.slots) {
+		return nil
+	}
+	x := c.slots[i]
+	if slot >= x.start && slot <= x.end {
+		return x.nodes
+	}
+	return nil
+}
+
+//------------------------------------------------------------------------------
+
+type clusterStateHolder struct {
+	load func(ctx context.Context) (*clusterState, error)
+
+	state     atomic.Value
+	reloading uint32 // atomic
+}
+
+func newClusterStateHolder(fn func(ctx context.Context) (*clusterState, error)) *clusterStateHolder {
+	return &clusterStateHolder{
+		load: fn,
+	}
+}
+
+func (c *clusterStateHolder) Reload(ctx context.Context) (*clusterState, error) {
+	state, err := c.load(ctx)
+	if err != nil {
+		return nil, err
+	}
+	c.state.Store(state)
+	return state, nil
+}
+
+func (c *clusterStateHolder) LazyReload(ctx context.Context) {
+	if !atomic.CompareAndSwapUint32(&c.reloading, 0, 1) {
+		return
+	}
+	go func() {
+		defer atomic.StoreUint32(&c.reloading, 0)
+
+		_, err := c.Reload(ctx)
+		if err != nil {
+			return
+		}
+		time.Sleep(200 * time.Millisecond)
+	}()
+}
+
+func (c *clusterStateHolder) Get(ctx context.Context) (*clusterState, error) {
+	v := c.state.Load()
+	if v != nil {
+		state := v.(*clusterState)
+		if time.Since(state.createdAt) > 10*time.Second {
+			c.LazyReload(ctx)
+		}
+		return state, nil
+	}
+	return c.Reload(ctx)
+}
+
+func (c *clusterStateHolder) ReloadOrGet(ctx context.Context) (*clusterState, error) {
+	state, err := c.Reload(ctx)
+	if err == nil {
+		return state, nil
+	}
+	return c.Get(ctx)
+}
+
+//------------------------------------------------------------------------------
+
+type clusterClient struct {
+	opt           *ClusterOptions
+	nodes         *clusterNodes
+	state         *clusterStateHolder //nolint:structcheck
+	cmdsInfoCache *cmdsInfoCache      //nolint:structcheck
+}
+
+// ClusterClient is a Redis Cluster client representing a pool of zero
+// or more underlying connections. It's safe for concurrent use by
+// multiple goroutines.
+type ClusterClient struct {
+	*clusterClient
+	cmdable
+	hooks
+	ctx context.Context
+}
+
+// NewClusterClient returns a Redis Cluster client as described in
+// http://redis.io/topics/cluster-spec.
+func NewClusterClient(opt *ClusterOptions) *ClusterClient {
+	opt.init()
+
+	c := &ClusterClient{
+		clusterClient: &clusterClient{
+			opt:   opt,
+			nodes: newClusterNodes(opt),
+		},
+		ctx: context.Background(),
+	}
+	c.state = newClusterStateHolder(c.loadState)
+	c.cmdsInfoCache = newCmdsInfoCache(c.cmdsInfo)
+	c.cmdable = c.Process
+
+	if opt.IdleCheckFrequency > 0 {
+		go c.reaper(opt.IdleCheckFrequency)
+	}
+
+	return c
+}
+
+func (c *ClusterClient) Context() context.Context {
+	return c.ctx
+}
+
+func (c *ClusterClient) WithContext(ctx context.Context) *ClusterClient {
+	if ctx == nil {
+		panic("nil context")
+	}
+	clone := *c
+	clone.cmdable = clone.Process
+	clone.hooks.lock()
+	clone.ctx = ctx
+	return &clone
+}
+
+// Options returns read-only Options that were used to create the client.
+func (c *ClusterClient) Options() *ClusterOptions {
+	return c.opt
+}
+
+// ReloadState reloads cluster state. If available it calls ClusterSlots func
+// to get cluster slots information.
+func (c *ClusterClient) ReloadState(ctx context.Context) {
+	c.state.LazyReload(ctx)
+}
+
+// Close closes the cluster client, releasing any open resources.
+//
+// It is rare to Close a ClusterClient, as the ClusterClient is meant
+// to be long-lived and shared between many goroutines.
+func (c *ClusterClient) Close() error {
+	return c.nodes.Close()
+}
+
+// Do creates a Cmd from the args and processes the cmd.
+func (c *ClusterClient) Do(ctx context.Context, args ...interface{}) *Cmd {
+	cmd := NewCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+func (c *ClusterClient) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.process)
+}
+
+func (c *ClusterClient) process(ctx context.Context, cmd Cmder) error {
+	cmdInfo := c.cmdInfo(cmd.Name())
+	slot := c.cmdSlot(cmd)
+
+	var node *clusterNode
+	var ask bool
+	var lastErr error
+	for attempt := 0; attempt <= c.opt.MaxRedirects; attempt++ {
+		if attempt > 0 {
+			if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+				return err
+			}
+		}
+
+		if node == nil {
+			var err error
+			node, err = c.cmdNode(ctx, cmdInfo, slot)
+			if err != nil {
+				return err
+			}
+		}
+
+		if ask {
+			pipe := node.Client.Pipeline()
+			_ = pipe.Process(ctx, NewCmd(ctx, "asking"))
+			_ = pipe.Process(ctx, cmd)
+			_, lastErr = pipe.Exec(ctx)
+			_ = pipe.Close()
+			ask = false
+		} else {
+			lastErr = node.Client.Process(ctx, cmd)
+		}
+
+		// If there is no error - we are done.
+		if lastErr == nil {
+			return nil
+		}
+		if isReadOnly := isReadOnlyError(lastErr); isReadOnly || lastErr == pool.ErrClosed {
+			if isReadOnly {
+				c.state.LazyReload(ctx)
+			}
+			node = nil
+			continue
+		}
+
+		// If slave is loading - pick another node.
+		if c.opt.ReadOnly && isLoadingError(lastErr) {
+			node.MarkAsFailing()
+			node = nil
+			continue
+		}
+
+		var moved bool
+		var addr string
+		moved, ask, addr = isMovedError(lastErr)
+		if moved || ask {
+			var err error
+			node, err = c.nodes.Get(addr)
+			if err != nil {
+				return err
+			}
+			continue
+		}
+
+		if shouldRetry(lastErr, cmd.readTimeout() == nil) {
+			// First retry the same node.
+			if attempt == 0 {
+				continue
+			}
+
+			// Second try another node.
+			node.MarkAsFailing()
+			node = nil
+			continue
+		}
+
+		return lastErr
+	}
+	return lastErr
+}
+
+// ForEachMaster concurrently calls the fn on each master node in the cluster.
+// It returns the first error if any.
+func (c *ClusterClient) ForEachMaster(
+	ctx context.Context,
+	fn func(ctx context.Context, client *Client) error,
+) error {
+	state, err := c.state.ReloadOrGet(ctx)
+	if err != nil {
+		return err
+	}
+
+	var wg sync.WaitGroup
+	errCh := make(chan error, 1)
+
+	for _, master := range state.Masters {
+		wg.Add(1)
+		go func(node *clusterNode) {
+			defer wg.Done()
+			err := fn(ctx, node.Client)
+			if err != nil {
+				select {
+				case errCh <- err:
+				default:
+				}
+			}
+		}(master)
+	}
+
+	wg.Wait()
+
+	select {
+	case err := <-errCh:
+		return err
+	default:
+		return nil
+	}
+}
+
+// ForEachSlave concurrently calls the fn on each slave node in the cluster.
+// It returns the first error if any.
+func (c *ClusterClient) ForEachSlave(
+	ctx context.Context,
+	fn func(ctx context.Context, client *Client) error,
+) error {
+	state, err := c.state.ReloadOrGet(ctx)
+	if err != nil {
+		return err
+	}
+
+	var wg sync.WaitGroup
+	errCh := make(chan error, 1)
+
+	for _, slave := range state.Slaves {
+		wg.Add(1)
+		go func(node *clusterNode) {
+			defer wg.Done()
+			err := fn(ctx, node.Client)
+			if err != nil {
+				select {
+				case errCh <- err:
+				default:
+				}
+			}
+		}(slave)
+	}
+
+	wg.Wait()
+
+	select {
+	case err := <-errCh:
+		return err
+	default:
+		return nil
+	}
+}
+
+// ForEachShard concurrently calls the fn on each known node in the cluster.
+// It returns the first error if any.
+func (c *ClusterClient) ForEachShard(
+	ctx context.Context,
+	fn func(ctx context.Context, client *Client) error,
+) error {
+	state, err := c.state.ReloadOrGet(ctx)
+	if err != nil {
+		return err
+	}
+
+	var wg sync.WaitGroup
+	errCh := make(chan error, 1)
+
+	worker := func(node *clusterNode) {
+		defer wg.Done()
+		err := fn(ctx, node.Client)
+		if err != nil {
+			select {
+			case errCh <- err:
+			default:
+			}
+		}
+	}
+
+	for _, node := range state.Masters {
+		wg.Add(1)
+		go worker(node)
+	}
+	for _, node := range state.Slaves {
+		wg.Add(1)
+		go worker(node)
+	}
+
+	wg.Wait()
+
+	select {
+	case err := <-errCh:
+		return err
+	default:
+		return nil
+	}
+}
+
+// PoolStats returns accumulated connection pool stats.
+func (c *ClusterClient) PoolStats() *PoolStats {
+	var acc PoolStats
+
+	state, _ := c.state.Get(context.TODO())
+	if state == nil {
+		return &acc
+	}
+
+	for _, node := range state.Masters {
+		s := node.Client.connPool.Stats()
+		acc.Hits += s.Hits
+		acc.Misses += s.Misses
+		acc.Timeouts += s.Timeouts
+
+		acc.TotalConns += s.TotalConns
+		acc.IdleConns += s.IdleConns
+		acc.StaleConns += s.StaleConns
+	}
+
+	for _, node := range state.Slaves {
+		s := node.Client.connPool.Stats()
+		acc.Hits += s.Hits
+		acc.Misses += s.Misses
+		acc.Timeouts += s.Timeouts
+
+		acc.TotalConns += s.TotalConns
+		acc.IdleConns += s.IdleConns
+		acc.StaleConns += s.StaleConns
+	}
+
+	return &acc
+}
+
+func (c *ClusterClient) loadState(ctx context.Context) (*clusterState, error) {
+	if c.opt.ClusterSlots != nil {
+		slots, err := c.opt.ClusterSlots(ctx)
+		if err != nil {
+			return nil, err
+		}
+		return newClusterState(c.nodes, slots, "")
+	}
+
+	addrs, err := c.nodes.Addrs()
+	if err != nil {
+		return nil, err
+	}
+
+	var firstErr error
+
+	for _, idx := range rand.Perm(len(addrs)) {
+		addr := addrs[idx]
+
+		node, err := c.nodes.Get(addr)
+		if err != nil {
+			if firstErr == nil {
+				firstErr = err
+			}
+			continue
+		}
+
+		slots, err := node.Client.ClusterSlots(ctx).Result()
+		if err != nil {
+			if firstErr == nil {
+				firstErr = err
+			}
+			continue
+		}
+
+		return newClusterState(c.nodes, slots, node.Client.opt.Addr)
+	}
+
+	/*
+	 * No node is connectable. It's possible that all nodes' IP has changed.
+	 * Clear activeAddrs to let client be able to re-connect using the initial
+	 * setting of the addresses (e.g. [redis-cluster-0:6379, redis-cluster-1:6379]),
+	 * which might have chance to resolve domain name and get updated IP address.
+	 */
+	c.nodes.mu.Lock()
+	c.nodes.activeAddrs = nil
+	c.nodes.mu.Unlock()
+
+	return nil, firstErr
+}
+
+// reaper closes idle connections to the cluster.
+func (c *ClusterClient) reaper(idleCheckFrequency time.Duration) {
+	ticker := time.NewTicker(idleCheckFrequency)
+	defer ticker.Stop()
+
+	for range ticker.C {
+		nodes, err := c.nodes.All()
+		if err != nil {
+			break
+		}
+
+		for _, node := range nodes {
+			_, err := node.Client.connPool.(*pool.ConnPool).ReapStaleConns()
+			if err != nil {
+				internal.Logger.Printf(c.Context(), "ReapStaleConns failed: %s", err)
+			}
+		}
+	}
+}
+
+func (c *ClusterClient) Pipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *ClusterClient) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipeline().Pipelined(ctx, fn)
+}
+
+func (c *ClusterClient) processPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, c._processPipeline)
+}
+
+func (c *ClusterClient) _processPipeline(ctx context.Context, cmds []Cmder) error {
+	cmdsMap := newCmdsMap()
+	err := c.mapCmdsByNode(ctx, cmdsMap, cmds)
+	if err != nil {
+		setCmdsErr(cmds, err)
+		return err
+	}
+
+	for attempt := 0; attempt <= c.opt.MaxRedirects; attempt++ {
+		if attempt > 0 {
+			if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+				setCmdsErr(cmds, err)
+				return err
+			}
+		}
+
+		failedCmds := newCmdsMap()
+		var wg sync.WaitGroup
+
+		for node, cmds := range cmdsMap.m {
+			wg.Add(1)
+			go func(node *clusterNode, cmds []Cmder) {
+				defer wg.Done()
+
+				err := c._processPipelineNode(ctx, node, cmds, failedCmds)
+				if err == nil {
+					return
+				}
+				if attempt < c.opt.MaxRedirects {
+					if err := c.mapCmdsByNode(ctx, failedCmds, cmds); err != nil {
+						setCmdsErr(cmds, err)
+					}
+				} else {
+					setCmdsErr(cmds, err)
+				}
+			}(node, cmds)
+		}
+
+		wg.Wait()
+		if len(failedCmds.m) == 0 {
+			break
+		}
+		cmdsMap = failedCmds
+	}
+
+	return cmdsFirstErr(cmds)
+}
+
+func (c *ClusterClient) mapCmdsByNode(ctx context.Context, cmdsMap *cmdsMap, cmds []Cmder) error {
+	state, err := c.state.Get(ctx)
+	if err != nil {
+		return err
+	}
+
+	if c.opt.ReadOnly && c.cmdsAreReadOnly(cmds) {
+		for _, cmd := range cmds {
+			slot := c.cmdSlot(cmd)
+			node, err := c.slotReadOnlyNode(state, slot)
+			if err != nil {
+				return err
+			}
+			cmdsMap.Add(node, cmd)
+		}
+		return nil
+	}
+
+	for _, cmd := range cmds {
+		slot := c.cmdSlot(cmd)
+		node, err := state.slotMasterNode(slot)
+		if err != nil {
+			return err
+		}
+		cmdsMap.Add(node, cmd)
+	}
+	return nil
+}
+
+func (c *ClusterClient) cmdsAreReadOnly(cmds []Cmder) bool {
+	for _, cmd := range cmds {
+		cmdInfo := c.cmdInfo(cmd.Name())
+		if cmdInfo == nil || !cmdInfo.ReadOnly {
+			return false
+		}
+	}
+	return true
+}
+
+func (c *ClusterClient) _processPipelineNode(
+	ctx context.Context, node *clusterNode, cmds []Cmder, failedCmds *cmdsMap,
+) error {
+	return node.Client.hooks.processPipeline(ctx, cmds, func(ctx context.Context, cmds []Cmder) error {
+		return node.Client.withConn(ctx, func(ctx context.Context, cn *pool.Conn) error {
+			err := cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+				return writeCmds(wr, cmds)
+			})
+			if err != nil {
+				return err
+			}
+
+			return cn.WithReader(ctx, c.opt.ReadTimeout, func(rd *proto.Reader) error {
+				return c.pipelineReadCmds(ctx, node, rd, cmds, failedCmds)
+			})
+		})
+	})
+}
+
+func (c *ClusterClient) pipelineReadCmds(
+	ctx context.Context,
+	node *clusterNode,
+	rd *proto.Reader,
+	cmds []Cmder,
+	failedCmds *cmdsMap,
+) error {
+	for _, cmd := range cmds {
+		err := cmd.readReply(rd)
+		cmd.SetErr(err)
+
+		if err == nil {
+			continue
+		}
+
+		if c.checkMovedErr(ctx, cmd, err, failedCmds) {
+			continue
+		}
+
+		if c.opt.ReadOnly && isLoadingError(err) {
+			node.MarkAsFailing()
+			return err
+		}
+		if isRedisError(err) {
+			continue
+		}
+		return err
+	}
+	return nil
+}
+
+func (c *ClusterClient) checkMovedErr(
+	ctx context.Context, cmd Cmder, err error, failedCmds *cmdsMap,
+) bool {
+	moved, ask, addr := isMovedError(err)
+	if !moved && !ask {
+		return false
+	}
+
+	node, err := c.nodes.Get(addr)
+	if err != nil {
+		return false
+	}
+
+	if moved {
+		c.state.LazyReload(ctx)
+		failedCmds.Add(node, cmd)
+		return true
+	}
+
+	if ask {
+		failedCmds.Add(node, NewCmd(ctx, "asking"), cmd)
+		return true
+	}
+
+	panic("not reached")
+}
+
+// TxPipeline acts like Pipeline, but wraps queued commands with MULTI/EXEC.
+func (c *ClusterClient) TxPipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processTxPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *ClusterClient) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.TxPipeline().Pipelined(ctx, fn)
+}
+
+func (c *ClusterClient) processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, c._processTxPipeline)
+}
+
+func (c *ClusterClient) _processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	state, err := c.state.Get(ctx)
+	if err != nil {
+		setCmdsErr(cmds, err)
+		return err
+	}
+
+	cmdsMap := c.mapCmdsBySlot(cmds)
+	for slot, cmds := range cmdsMap {
+		node, err := state.slotMasterNode(slot)
+		if err != nil {
+			setCmdsErr(cmds, err)
+			continue
+		}
+
+		cmdsMap := map[*clusterNode][]Cmder{node: cmds}
+		for attempt := 0; attempt <= c.opt.MaxRedirects; attempt++ {
+			if attempt > 0 {
+				if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+					setCmdsErr(cmds, err)
+					return err
+				}
+			}
+
+			failedCmds := newCmdsMap()
+			var wg sync.WaitGroup
+
+			for node, cmds := range cmdsMap {
+				wg.Add(1)
+				go func(node *clusterNode, cmds []Cmder) {
+					defer wg.Done()
+
+					err := c._processTxPipelineNode(ctx, node, cmds, failedCmds)
+					if err == nil {
+						return
+					}
+					if attempt < c.opt.MaxRedirects {
+						if err := c.mapCmdsByNode(ctx, failedCmds, cmds); err != nil {
+							setCmdsErr(cmds, err)
+						}
+					} else {
+						setCmdsErr(cmds, err)
+					}
+				}(node, cmds)
+			}
+
+			wg.Wait()
+			if len(failedCmds.m) == 0 {
+				break
+			}
+			cmdsMap = failedCmds.m
+		}
+	}
+
+	return cmdsFirstErr(cmds)
+}
+
+func (c *ClusterClient) mapCmdsBySlot(cmds []Cmder) map[int][]Cmder {
+	cmdsMap := make(map[int][]Cmder)
+	for _, cmd := range cmds {
+		slot := c.cmdSlot(cmd)
+		cmdsMap[slot] = append(cmdsMap[slot], cmd)
+	}
+	return cmdsMap
+}
+
+func (c *ClusterClient) _processTxPipelineNode(
+	ctx context.Context, node *clusterNode, cmds []Cmder, failedCmds *cmdsMap,
+) error {
+	return node.Client.hooks.processTxPipeline(ctx, cmds, func(ctx context.Context, cmds []Cmder) error {
+		return node.Client.withConn(ctx, func(ctx context.Context, cn *pool.Conn) error {
+			err := cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+				return writeCmds(wr, cmds)
+			})
+			if err != nil {
+				return err
+			}
+
+			return cn.WithReader(ctx, c.opt.ReadTimeout, func(rd *proto.Reader) error {
+				statusCmd := cmds[0].(*StatusCmd)
+				// Trim multi and exec.
+				cmds = cmds[1 : len(cmds)-1]
+
+				err := c.txPipelineReadQueued(ctx, rd, statusCmd, cmds, failedCmds)
+				if err != nil {
+					moved, ask, addr := isMovedError(err)
+					if moved || ask {
+						return c.cmdsMoved(ctx, cmds, moved, ask, addr, failedCmds)
+					}
+					return err
+				}
+
+				return pipelineReadCmds(rd, cmds)
+			})
+		})
+	})
+}
+
+func (c *ClusterClient) txPipelineReadQueued(
+	ctx context.Context,
+	rd *proto.Reader,
+	statusCmd *StatusCmd,
+	cmds []Cmder,
+	failedCmds *cmdsMap,
+) error {
+	// Parse queued replies.
+	if err := statusCmd.readReply(rd); err != nil {
+		return err
+	}
+
+	for _, cmd := range cmds {
+		err := statusCmd.readReply(rd)
+		if err == nil || c.checkMovedErr(ctx, cmd, err, failedCmds) || isRedisError(err) {
+			continue
+		}
+		return err
+	}
+
+	// Parse number of replies.
+	line, err := rd.ReadLine()
+	if err != nil {
+		if err == Nil {
+			err = TxFailedErr
+		}
+		return err
+	}
+
+	switch line[0] {
+	case proto.ErrorReply:
+		return proto.ParseErrorReply(line)
+	case proto.ArrayReply:
+		// ok
+	default:
+		return fmt.Errorf("redis: expected '*', but got line %q", line)
+	}
+
+	return nil
+}
+
+func (c *ClusterClient) cmdsMoved(
+	ctx context.Context, cmds []Cmder,
+	moved, ask bool,
+	addr string,
+	failedCmds *cmdsMap,
+) error {
+	node, err := c.nodes.Get(addr)
+	if err != nil {
+		return err
+	}
+
+	if moved {
+		c.state.LazyReload(ctx)
+		for _, cmd := range cmds {
+			failedCmds.Add(node, cmd)
+		}
+		return nil
+	}
+
+	if ask {
+		for _, cmd := range cmds {
+			failedCmds.Add(node, NewCmd(ctx, "asking"), cmd)
+		}
+		return nil
+	}
+
+	return nil
+}
+
+func (c *ClusterClient) Watch(ctx context.Context, fn func(*Tx) error, keys ...string) error {
+	if len(keys) == 0 {
+		return fmt.Errorf("redis: Watch requires at least one key")
+	}
+
+	slot := hashtag.Slot(keys[0])
+	for _, key := range keys[1:] {
+		if hashtag.Slot(key) != slot {
+			err := fmt.Errorf("redis: Watch requires all keys to be in the same slot")
+			return err
+		}
+	}
+
+	node, err := c.slotMasterNode(ctx, slot)
+	if err != nil {
+		return err
+	}
+
+	for attempt := 0; attempt <= c.opt.MaxRedirects; attempt++ {
+		if attempt > 0 {
+			if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+				return err
+			}
+		}
+
+		err = node.Client.Watch(ctx, fn, keys...)
+		if err == nil {
+			break
+		}
+
+		moved, ask, addr := isMovedError(err)
+		if moved || ask {
+			node, err = c.nodes.Get(addr)
+			if err != nil {
+				return err
+			}
+			continue
+		}
+
+		if isReadOnly := isReadOnlyError(err); isReadOnly || err == pool.ErrClosed {
+			if isReadOnly {
+				c.state.LazyReload(ctx)
+			}
+			node, err = c.slotMasterNode(ctx, slot)
+			if err != nil {
+				return err
+			}
+			continue
+		}
+
+		if shouldRetry(err, true) {
+			continue
+		}
+
+		return err
+	}
+
+	return err
+}
+
+func (c *ClusterClient) pubSub() *PubSub {
+	var node *clusterNode
+	pubsub := &PubSub{
+		opt: c.opt.clientOptions(),
+
+		newConn: func(ctx context.Context, channels []string) (*pool.Conn, error) {
+			if node != nil {
+				panic("node != nil")
+			}
+
+			var err error
+			if len(channels) > 0 {
+				slot := hashtag.Slot(channels[0])
+				node, err = c.slotMasterNode(ctx, slot)
+			} else {
+				node, err = c.nodes.Random()
+			}
+			if err != nil {
+				return nil, err
+			}
+
+			cn, err := node.Client.newConn(context.TODO())
+			if err != nil {
+				node = nil
+
+				return nil, err
+			}
+
+			return cn, nil
+		},
+		closeConn: func(cn *pool.Conn) error {
+			err := node.Client.connPool.CloseConn(cn)
+			node = nil
+			return err
+		},
+	}
+	pubsub.init()
+
+	return pubsub
+}
+
+// Subscribe subscribes the client to the specified channels.
+// Channels can be omitted to create empty subscription.
+func (c *ClusterClient) Subscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.Subscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+// PSubscribe subscribes the client to the given patterns.
+// Patterns can be omitted to create empty subscription.
+func (c *ClusterClient) PSubscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.PSubscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+func (c *ClusterClient) retryBackoff(attempt int) time.Duration {
+	return internal.RetryBackoff(attempt, c.opt.MinRetryBackoff, c.opt.MaxRetryBackoff)
+}
+
+func (c *ClusterClient) cmdsInfo(ctx context.Context) (map[string]*CommandInfo, error) {
+	// Try 3 random nodes.
+	const nodeLimit = 3
+
+	addrs, err := c.nodes.Addrs()
+	if err != nil {
+		return nil, err
+	}
+
+	var firstErr error
+
+	perm := rand.Perm(len(addrs))
+	if len(perm) > nodeLimit {
+		perm = perm[:nodeLimit]
+	}
+
+	for _, idx := range perm {
+		addr := addrs[idx]
+
+		node, err := c.nodes.Get(addr)
+		if err != nil {
+			if firstErr == nil {
+				firstErr = err
+			}
+			continue
+		}
+
+		info, err := node.Client.Command(ctx).Result()
+		if err == nil {
+			return info, nil
+		}
+		if firstErr == nil {
+			firstErr = err
+		}
+	}
+
+	if firstErr == nil {
+		panic("not reached")
+	}
+	return nil, firstErr
+}
+
+func (c *ClusterClient) cmdInfo(name string) *CommandInfo {
+	cmdsInfo, err := c.cmdsInfoCache.Get(c.ctx)
+	if err != nil {
+		return nil
+	}
+
+	info := cmdsInfo[name]
+	if info == nil {
+		internal.Logger.Printf(c.Context(), "info for cmd=%s not found", name)
+	}
+	return info
+}
+
+func (c *ClusterClient) cmdSlot(cmd Cmder) int {
+	args := cmd.Args()
+	if args[0] == "cluster" && args[1] == "getkeysinslot" {
+		return args[2].(int)
+	}
+
+	cmdInfo := c.cmdInfo(cmd.Name())
+	return cmdSlot(cmd, cmdFirstKeyPos(cmd, cmdInfo))
+}
+
+func cmdSlot(cmd Cmder, pos int) int {
+	if pos == 0 {
+		return hashtag.RandomSlot()
+	}
+	firstKey := cmd.stringArg(pos)
+	return hashtag.Slot(firstKey)
+}
+
+func (c *ClusterClient) cmdNode(
+	ctx context.Context,
+	cmdInfo *CommandInfo,
+	slot int,
+) (*clusterNode, error) {
+	state, err := c.state.Get(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	if (c.opt.RouteByLatency || c.opt.RouteRandomly) && cmdInfo != nil && cmdInfo.ReadOnly {
+		return c.slotReadOnlyNode(state, slot)
+	}
+	return state.slotMasterNode(slot)
+}
+
+func (c *clusterClient) slotReadOnlyNode(state *clusterState, slot int) (*clusterNode, error) {
+	if c.opt.RouteByLatency {
+		return state.slotClosestNode(slot)
+	}
+	if c.opt.RouteRandomly {
+		return state.slotRandomNode(slot)
+	}
+	return state.slotSlaveNode(slot)
+}
+
+func (c *ClusterClient) slotMasterNode(ctx context.Context, slot int) (*clusterNode, error) {
+	state, err := c.state.Get(ctx)
+	if err != nil {
+		return nil, err
+	}
+	return state.slotMasterNode(slot)
+}
+
+func appendUniqueNode(nodes []*clusterNode, node *clusterNode) []*clusterNode {
+	for _, n := range nodes {
+		if n == node {
+			return nodes
+		}
+	}
+	return append(nodes, node)
+}
+
+func appendIfNotExists(ss []string, es ...string) []string {
+loop:
+	for _, e := range es {
+		for _, s := range ss {
+			if s == e {
+				continue loop
+			}
+		}
+		ss = append(ss, e)
+	}
+	return ss
+}
+
+//------------------------------------------------------------------------------
+
+type cmdsMap struct {
+	mu sync.Mutex
+	m  map[*clusterNode][]Cmder
+}
+
+func newCmdsMap() *cmdsMap {
+	return &cmdsMap{
+		m: make(map[*clusterNode][]Cmder),
+	}
+}
+
+func (m *cmdsMap) Add(node *clusterNode, cmds ...Cmder) {
+	m.mu.Lock()
+	m.m[node] = append(m.m[node], cmds...)
+	m.mu.Unlock()
+}
diff --git a/vendor/github.com/go-redis/redis/v8/cluster_commands.go b/vendor/github.com/go-redis/redis/v8/cluster_commands.go
new file mode 100644
index 0000000..1f0bae0
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/cluster_commands.go
@@ -0,0 +1,25 @@
+package redis
+
+import (
+	"context"
+	"sync/atomic"
+)
+
+func (c *ClusterClient) DBSize(ctx context.Context) *IntCmd {
+	cmd := NewIntCmd(ctx, "dbsize")
+	var size int64
+	err := c.ForEachMaster(ctx, func(ctx context.Context, master *Client) error {
+		n, err := master.DBSize(ctx).Result()
+		if err != nil {
+			return err
+		}
+		atomic.AddInt64(&size, n)
+		return nil
+	})
+	if err != nil {
+		cmd.SetErr(err)
+		return cmd
+	}
+	cmd.val = size
+	return cmd
+}
diff --git a/vendor/github.com/go-redis/redis/v8/command.go b/vendor/github.com/go-redis/redis/v8/command.go
new file mode 100644
index 0000000..5dd5533
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/command.go
@@ -0,0 +1,2396 @@
+package redis
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"strconv"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/proto"
+	"github.com/go-redis/redis/v8/internal/util"
+)
+
+type Cmder interface {
+	Name() string
+	FullName() string
+	Args() []interface{}
+	String() string
+	stringArg(int) string
+	firstKeyPos() int8
+	setFirstKeyPos(int8)
+
+	readTimeout() *time.Duration
+	readReply(rd *proto.Reader) error
+
+	SetErr(error)
+	Err() error
+}
+
+func setCmdsErr(cmds []Cmder, e error) {
+	for _, cmd := range cmds {
+		if cmd.Err() == nil {
+			cmd.SetErr(e)
+		}
+	}
+}
+
+func cmdsFirstErr(cmds []Cmder) error {
+	for _, cmd := range cmds {
+		if err := cmd.Err(); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func writeCmds(wr *proto.Writer, cmds []Cmder) error {
+	for _, cmd := range cmds {
+		if err := writeCmd(wr, cmd); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func writeCmd(wr *proto.Writer, cmd Cmder) error {
+	return wr.WriteArgs(cmd.Args())
+}
+
+func cmdFirstKeyPos(cmd Cmder, info *CommandInfo) int {
+	if pos := cmd.firstKeyPos(); pos != 0 {
+		return int(pos)
+	}
+
+	switch cmd.Name() {
+	case "eval", "evalsha":
+		if cmd.stringArg(2) != "0" {
+			return 3
+		}
+
+		return 0
+	case "publish":
+		return 1
+	case "memory":
+		// https://github.com/redis/redis/issues/7493
+		if cmd.stringArg(1) == "usage" {
+			return 2
+		}
+	}
+
+	if info != nil {
+		return int(info.FirstKeyPos)
+	}
+	return 0
+}
+
+func cmdString(cmd Cmder, val interface{}) string {
+	b := make([]byte, 0, 64)
+
+	for i, arg := range cmd.Args() {
+		if i > 0 {
+			b = append(b, ' ')
+		}
+		b = internal.AppendArg(b, arg)
+	}
+
+	if err := cmd.Err(); err != nil {
+		b = append(b, ": "...)
+		b = append(b, err.Error()...)
+	} else if val != nil {
+		b = append(b, ": "...)
+		b = internal.AppendArg(b, val)
+	}
+
+	return internal.String(b)
+}
+
+//------------------------------------------------------------------------------
+
+type baseCmd struct {
+	ctx    context.Context
+	args   []interface{}
+	err    error
+	keyPos int8
+
+	_readTimeout *time.Duration
+}
+
+var _ Cmder = (*Cmd)(nil)
+
+func (cmd *baseCmd) Name() string {
+	if len(cmd.args) == 0 {
+		return ""
+	}
+	// Cmd name must be lower cased.
+	return internal.ToLower(cmd.stringArg(0))
+}
+
+func (cmd *baseCmd) FullName() string {
+	switch name := cmd.Name(); name {
+	case "cluster", "command":
+		if len(cmd.args) == 1 {
+			return name
+		}
+		if s2, ok := cmd.args[1].(string); ok {
+			return name + " " + s2
+		}
+		return name
+	default:
+		return name
+	}
+}
+
+func (cmd *baseCmd) Args() []interface{} {
+	return cmd.args
+}
+
+func (cmd *baseCmd) stringArg(pos int) string {
+	if pos < 0 || pos >= len(cmd.args) {
+		return ""
+	}
+	s, _ := cmd.args[pos].(string)
+	return s
+}
+
+func (cmd *baseCmd) firstKeyPos() int8 {
+	return cmd.keyPos
+}
+
+func (cmd *baseCmd) setFirstKeyPos(keyPos int8) {
+	cmd.keyPos = keyPos
+}
+
+func (cmd *baseCmd) SetErr(e error) {
+	cmd.err = e
+}
+
+func (cmd *baseCmd) Err() error {
+	return cmd.err
+}
+
+func (cmd *baseCmd) readTimeout() *time.Duration {
+	return cmd._readTimeout
+}
+
+func (cmd *baseCmd) setReadTimeout(d time.Duration) {
+	cmd._readTimeout = &d
+}
+
+//------------------------------------------------------------------------------
+
+type Cmd struct {
+	baseCmd
+
+	val interface{}
+}
+
+func NewCmd(ctx context.Context, args ...interface{}) *Cmd {
+	return &Cmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *Cmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *Cmd) Val() interface{} {
+	return cmd.val
+}
+
+func (cmd *Cmd) Result() (interface{}, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *Cmd) Text() (string, error) {
+	if cmd.err != nil {
+		return "", cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case string:
+		return val, nil
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for String", val)
+		return "", err
+	}
+}
+
+func (cmd *Cmd) Int() (int, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return int(val), nil
+	case string:
+		return strconv.Atoi(val)
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Int", val)
+		return 0, err
+	}
+}
+
+func (cmd *Cmd) Int64() (int64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return val, nil
+	case string:
+		return strconv.ParseInt(val, 10, 64)
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Int64", val)
+		return 0, err
+	}
+}
+
+func (cmd *Cmd) Uint64() (uint64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return uint64(val), nil
+	case string:
+		return strconv.ParseUint(val, 10, 64)
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Uint64", val)
+		return 0, err
+	}
+}
+
+func (cmd *Cmd) Float32() (float32, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return float32(val), nil
+	case string:
+		f, err := strconv.ParseFloat(val, 32)
+		if err != nil {
+			return 0, err
+		}
+		return float32(f), nil
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Float32", val)
+		return 0, err
+	}
+}
+
+func (cmd *Cmd) Float64() (float64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return float64(val), nil
+	case string:
+		return strconv.ParseFloat(val, 64)
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Float64", val)
+		return 0, err
+	}
+}
+
+func (cmd *Cmd) Bool() (bool, error) {
+	if cmd.err != nil {
+		return false, cmd.err
+	}
+	switch val := cmd.val.(type) {
+	case int64:
+		return val != 0, nil
+	case string:
+		return strconv.ParseBool(val)
+	default:
+		err := fmt.Errorf("redis: unexpected type=%T for Bool", val)
+		return false, err
+	}
+}
+
+func (cmd *Cmd) readReply(rd *proto.Reader) (err error) {
+	cmd.val, err = rd.ReadReply(sliceParser)
+	return err
+}
+
+// sliceParser implements proto.MultiBulkParse.
+func sliceParser(rd *proto.Reader, n int64) (interface{}, error) {
+	vals := make([]interface{}, n)
+	for i := 0; i < len(vals); i++ {
+		v, err := rd.ReadReply(sliceParser)
+		if err != nil {
+			if err == Nil {
+				vals[i] = nil
+				continue
+			}
+			if err, ok := err.(proto.RedisError); ok {
+				vals[i] = err
+				continue
+			}
+			return nil, err
+		}
+		vals[i] = v
+	}
+	return vals, nil
+}
+
+//------------------------------------------------------------------------------
+
+type SliceCmd struct {
+	baseCmd
+
+	val []interface{}
+}
+
+var _ Cmder = (*SliceCmd)(nil)
+
+func NewSliceCmd(ctx context.Context, args ...interface{}) *SliceCmd {
+	return &SliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *SliceCmd) Val() []interface{} {
+	return cmd.val
+}
+
+func (cmd *SliceCmd) Result() ([]interface{}, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *SliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *SliceCmd) readReply(rd *proto.Reader) error {
+	v, err := rd.ReadArrayReply(sliceParser)
+	if err != nil {
+		return err
+	}
+	cmd.val = v.([]interface{})
+	return nil
+}
+
+//------------------------------------------------------------------------------
+
+type StatusCmd struct {
+	baseCmd
+
+	val string
+}
+
+var _ Cmder = (*StatusCmd)(nil)
+
+func NewStatusCmd(ctx context.Context, args ...interface{}) *StatusCmd {
+	return &StatusCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StatusCmd) Val() string {
+	return cmd.val
+}
+
+func (cmd *StatusCmd) Result() (string, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *StatusCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StatusCmd) readReply(rd *proto.Reader) (err error) {
+	cmd.val, err = rd.ReadString()
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type IntCmd struct {
+	baseCmd
+
+	val int64
+}
+
+var _ Cmder = (*IntCmd)(nil)
+
+func NewIntCmd(ctx context.Context, args ...interface{}) *IntCmd {
+	return &IntCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *IntCmd) Val() int64 {
+	return cmd.val
+}
+
+func (cmd *IntCmd) Result() (int64, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *IntCmd) Uint64() (uint64, error) {
+	return uint64(cmd.val), cmd.err
+}
+
+func (cmd *IntCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *IntCmd) readReply(rd *proto.Reader) (err error) {
+	cmd.val, err = rd.ReadIntReply()
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type IntSliceCmd struct {
+	baseCmd
+
+	val []int64
+}
+
+var _ Cmder = (*IntSliceCmd)(nil)
+
+func NewIntSliceCmd(ctx context.Context, args ...interface{}) *IntSliceCmd {
+	return &IntSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *IntSliceCmd) Val() []int64 {
+	return cmd.val
+}
+
+func (cmd *IntSliceCmd) Result() ([]int64, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *IntSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *IntSliceCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]int64, n)
+		for i := 0; i < len(cmd.val); i++ {
+			num, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+			cmd.val[i] = num
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type DurationCmd struct {
+	baseCmd
+
+	val       time.Duration
+	precision time.Duration
+}
+
+var _ Cmder = (*DurationCmd)(nil)
+
+func NewDurationCmd(ctx context.Context, precision time.Duration, args ...interface{}) *DurationCmd {
+	return &DurationCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+		precision: precision,
+	}
+}
+
+func (cmd *DurationCmd) Val() time.Duration {
+	return cmd.val
+}
+
+func (cmd *DurationCmd) Result() (time.Duration, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *DurationCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *DurationCmd) readReply(rd *proto.Reader) error {
+	n, err := rd.ReadIntReply()
+	if err != nil {
+		return err
+	}
+	switch n {
+	// -2 if the key does not exist
+	// -1 if the key exists but has no associated expire
+	case -2, -1:
+		cmd.val = time.Duration(n)
+	default:
+		cmd.val = time.Duration(n) * cmd.precision
+	}
+	return nil
+}
+
+//------------------------------------------------------------------------------
+
+type TimeCmd struct {
+	baseCmd
+
+	val time.Time
+}
+
+var _ Cmder = (*TimeCmd)(nil)
+
+func NewTimeCmd(ctx context.Context, args ...interface{}) *TimeCmd {
+	return &TimeCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *TimeCmd) Val() time.Time {
+	return cmd.val
+}
+
+func (cmd *TimeCmd) Result() (time.Time, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *TimeCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *TimeCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		if n != 2 {
+			return nil, fmt.Errorf("got %d elements, expected 2", n)
+		}
+
+		sec, err := rd.ReadInt()
+		if err != nil {
+			return nil, err
+		}
+
+		microsec, err := rd.ReadInt()
+		if err != nil {
+			return nil, err
+		}
+
+		cmd.val = time.Unix(sec, microsec*1000)
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type BoolCmd struct {
+	baseCmd
+
+	val bool
+}
+
+var _ Cmder = (*BoolCmd)(nil)
+
+func NewBoolCmd(ctx context.Context, args ...interface{}) *BoolCmd {
+	return &BoolCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *BoolCmd) Val() bool {
+	return cmd.val
+}
+
+func (cmd *BoolCmd) Result() (bool, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *BoolCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *BoolCmd) readReply(rd *proto.Reader) error {
+	v, err := rd.ReadReply(nil)
+	// `SET key value NX` returns nil when key already exists. But
+	// `SETNX key value` returns bool (0/1). So convert nil to bool.
+	if err == Nil {
+		cmd.val = false
+		return nil
+	}
+	if err != nil {
+		return err
+	}
+	switch v := v.(type) {
+	case int64:
+		cmd.val = v == 1
+		return nil
+	case string:
+		cmd.val = v == "OK"
+		return nil
+	default:
+		return fmt.Errorf("got %T, wanted int64 or string", v)
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type StringCmd struct {
+	baseCmd
+
+	val string
+}
+
+var _ Cmder = (*StringCmd)(nil)
+
+func NewStringCmd(ctx context.Context, args ...interface{}) *StringCmd {
+	return &StringCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StringCmd) Val() string {
+	return cmd.val
+}
+
+func (cmd *StringCmd) Result() (string, error) {
+	return cmd.Val(), cmd.err
+}
+
+func (cmd *StringCmd) Bytes() ([]byte, error) {
+	return util.StringToBytes(cmd.val), cmd.err
+}
+
+func (cmd *StringCmd) Int() (int, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	return strconv.Atoi(cmd.Val())
+}
+
+func (cmd *StringCmd) Int64() (int64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	return strconv.ParseInt(cmd.Val(), 10, 64)
+}
+
+func (cmd *StringCmd) Uint64() (uint64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	return strconv.ParseUint(cmd.Val(), 10, 64)
+}
+
+func (cmd *StringCmd) Float32() (float32, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	f, err := strconv.ParseFloat(cmd.Val(), 32)
+	if err != nil {
+		return 0, err
+	}
+	return float32(f), nil
+}
+
+func (cmd *StringCmd) Float64() (float64, error) {
+	if cmd.err != nil {
+		return 0, cmd.err
+	}
+	return strconv.ParseFloat(cmd.Val(), 64)
+}
+
+func (cmd *StringCmd) Time() (time.Time, error) {
+	if cmd.err != nil {
+		return time.Time{}, cmd.err
+	}
+	return time.Parse(time.RFC3339Nano, cmd.Val())
+}
+
+func (cmd *StringCmd) Scan(val interface{}) error {
+	if cmd.err != nil {
+		return cmd.err
+	}
+	return proto.Scan([]byte(cmd.val), val)
+}
+
+func (cmd *StringCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StringCmd) readReply(rd *proto.Reader) (err error) {
+	cmd.val, err = rd.ReadString()
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type FloatCmd struct {
+	baseCmd
+
+	val float64
+}
+
+var _ Cmder = (*FloatCmd)(nil)
+
+func NewFloatCmd(ctx context.Context, args ...interface{}) *FloatCmd {
+	return &FloatCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *FloatCmd) Val() float64 {
+	return cmd.val
+}
+
+func (cmd *FloatCmd) Result() (float64, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *FloatCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *FloatCmd) readReply(rd *proto.Reader) (err error) {
+	cmd.val, err = rd.ReadFloatReply()
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type StringSliceCmd struct {
+	baseCmd
+
+	val []string
+}
+
+var _ Cmder = (*StringSliceCmd)(nil)
+
+func NewStringSliceCmd(ctx context.Context, args ...interface{}) *StringSliceCmd {
+	return &StringSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StringSliceCmd) Val() []string {
+	return cmd.val
+}
+
+func (cmd *StringSliceCmd) Result() ([]string, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *StringSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StringSliceCmd) ScanSlice(container interface{}) error {
+	return proto.ScanSlice(cmd.Val(), container)
+}
+
+func (cmd *StringSliceCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]string, n)
+		for i := 0; i < len(cmd.val); i++ {
+			switch s, err := rd.ReadString(); {
+			case err == Nil:
+				cmd.val[i] = ""
+			case err != nil:
+				return nil, err
+			default:
+				cmd.val[i] = s
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type BoolSliceCmd struct {
+	baseCmd
+
+	val []bool
+}
+
+var _ Cmder = (*BoolSliceCmd)(nil)
+
+func NewBoolSliceCmd(ctx context.Context, args ...interface{}) *BoolSliceCmd {
+	return &BoolSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *BoolSliceCmd) Val() []bool {
+	return cmd.val
+}
+
+func (cmd *BoolSliceCmd) Result() ([]bool, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *BoolSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *BoolSliceCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]bool, n)
+		for i := 0; i < len(cmd.val); i++ {
+			n, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+			cmd.val[i] = n == 1
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type StringStringMapCmd struct {
+	baseCmd
+
+	val map[string]string
+}
+
+var _ Cmder = (*StringStringMapCmd)(nil)
+
+func NewStringStringMapCmd(ctx context.Context, args ...interface{}) *StringStringMapCmd {
+	return &StringStringMapCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StringStringMapCmd) Val() map[string]string {
+	return cmd.val
+}
+
+func (cmd *StringStringMapCmd) Result() (map[string]string, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *StringStringMapCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StringStringMapCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make(map[string]string, n/2)
+		for i := int64(0); i < n; i += 2 {
+			key, err := rd.ReadString()
+			if err != nil {
+				return nil, err
+			}
+
+			value, err := rd.ReadString()
+			if err != nil {
+				return nil, err
+			}
+
+			cmd.val[key] = value
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type StringIntMapCmd struct {
+	baseCmd
+
+	val map[string]int64
+}
+
+var _ Cmder = (*StringIntMapCmd)(nil)
+
+func NewStringIntMapCmd(ctx context.Context, args ...interface{}) *StringIntMapCmd {
+	return &StringIntMapCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StringIntMapCmd) Val() map[string]int64 {
+	return cmd.val
+}
+
+func (cmd *StringIntMapCmd) Result() (map[string]int64, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *StringIntMapCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StringIntMapCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make(map[string]int64, n/2)
+		for i := int64(0); i < n; i += 2 {
+			key, err := rd.ReadString()
+			if err != nil {
+				return nil, err
+			}
+
+			n, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+
+			cmd.val[key] = n
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type StringStructMapCmd struct {
+	baseCmd
+
+	val map[string]struct{}
+}
+
+var _ Cmder = (*StringStructMapCmd)(nil)
+
+func NewStringStructMapCmd(ctx context.Context, args ...interface{}) *StringStructMapCmd {
+	return &StringStructMapCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *StringStructMapCmd) Val() map[string]struct{} {
+	return cmd.val
+}
+
+func (cmd *StringStructMapCmd) Result() (map[string]struct{}, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *StringStructMapCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *StringStructMapCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make(map[string]struct{}, n)
+		for i := int64(0); i < n; i++ {
+			key, err := rd.ReadString()
+			if err != nil {
+				return nil, err
+			}
+			cmd.val[key] = struct{}{}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type XMessage struct {
+	ID     string
+	Values map[string]interface{}
+}
+
+type XMessageSliceCmd struct {
+	baseCmd
+
+	val []XMessage
+}
+
+var _ Cmder = (*XMessageSliceCmd)(nil)
+
+func NewXMessageSliceCmd(ctx context.Context, args ...interface{}) *XMessageSliceCmd {
+	return &XMessageSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *XMessageSliceCmd) Val() []XMessage {
+	return cmd.val
+}
+
+func (cmd *XMessageSliceCmd) Result() ([]XMessage, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XMessageSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XMessageSliceCmd) readReply(rd *proto.Reader) error {
+	var err error
+	cmd.val, err = readXMessageSlice(rd)
+	return err
+}
+
+func readXMessageSlice(rd *proto.Reader) ([]XMessage, error) {
+	n, err := rd.ReadArrayLen()
+	if err != nil {
+		return nil, err
+	}
+
+	msgs := make([]XMessage, n)
+	for i := 0; i < n; i++ {
+		var err error
+		msgs[i], err = readXMessage(rd)
+		if err != nil {
+			return nil, err
+		}
+	}
+	return msgs, nil
+}
+
+func readXMessage(rd *proto.Reader) (XMessage, error) {
+	n, err := rd.ReadArrayLen()
+	if err != nil {
+		return XMessage{}, err
+	}
+	if n != 2 {
+		return XMessage{}, fmt.Errorf("got %d, wanted 2", n)
+	}
+
+	id, err := rd.ReadString()
+	if err != nil {
+		return XMessage{}, err
+	}
+
+	var values map[string]interface{}
+
+	v, err := rd.ReadArrayReply(stringInterfaceMapParser)
+	if err != nil {
+		if err != proto.Nil {
+			return XMessage{}, err
+		}
+	} else {
+		values = v.(map[string]interface{})
+	}
+
+	return XMessage{
+		ID:     id,
+		Values: values,
+	}, nil
+}
+
+// stringInterfaceMapParser implements proto.MultiBulkParse.
+func stringInterfaceMapParser(rd *proto.Reader, n int64) (interface{}, error) {
+	m := make(map[string]interface{}, n/2)
+	for i := int64(0); i < n; i += 2 {
+		key, err := rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+
+		value, err := rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+
+		m[key] = value
+	}
+	return m, nil
+}
+
+//------------------------------------------------------------------------------
+
+type XStream struct {
+	Stream   string
+	Messages []XMessage
+}
+
+type XStreamSliceCmd struct {
+	baseCmd
+
+	val []XStream
+}
+
+var _ Cmder = (*XStreamSliceCmd)(nil)
+
+func NewXStreamSliceCmd(ctx context.Context, args ...interface{}) *XStreamSliceCmd {
+	return &XStreamSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *XStreamSliceCmd) Val() []XStream {
+	return cmd.val
+}
+
+func (cmd *XStreamSliceCmd) Result() ([]XStream, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XStreamSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XStreamSliceCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]XStream, n)
+		for i := 0; i < len(cmd.val); i++ {
+			i := i
+			_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+				if n != 2 {
+					return nil, fmt.Errorf("got %d, wanted 2", n)
+				}
+
+				stream, err := rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+
+				msgs, err := readXMessageSlice(rd)
+				if err != nil {
+					return nil, err
+				}
+
+				cmd.val[i] = XStream{
+					Stream:   stream,
+					Messages: msgs,
+				}
+				return nil, nil
+			})
+			if err != nil {
+				return nil, err
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type XPending struct {
+	Count     int64
+	Lower     string
+	Higher    string
+	Consumers map[string]int64
+}
+
+type XPendingCmd struct {
+	baseCmd
+	val *XPending
+}
+
+var _ Cmder = (*XPendingCmd)(nil)
+
+func NewXPendingCmd(ctx context.Context, args ...interface{}) *XPendingCmd {
+	return &XPendingCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *XPendingCmd) Val() *XPending {
+	return cmd.val
+}
+
+func (cmd *XPendingCmd) Result() (*XPending, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XPendingCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XPendingCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		if n != 4 {
+			return nil, fmt.Errorf("got %d, wanted 4", n)
+		}
+
+		count, err := rd.ReadIntReply()
+		if err != nil {
+			return nil, err
+		}
+
+		lower, err := rd.ReadString()
+		if err != nil && err != Nil {
+			return nil, err
+		}
+
+		higher, err := rd.ReadString()
+		if err != nil && err != Nil {
+			return nil, err
+		}
+
+		cmd.val = &XPending{
+			Count:  count,
+			Lower:  lower,
+			Higher: higher,
+		}
+		_, err = rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+			for i := int64(0); i < n; i++ {
+				_, err = rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+					if n != 2 {
+						return nil, fmt.Errorf("got %d, wanted 2", n)
+					}
+
+					consumerName, err := rd.ReadString()
+					if err != nil {
+						return nil, err
+					}
+
+					consumerPending, err := rd.ReadInt()
+					if err != nil {
+						return nil, err
+					}
+
+					if cmd.val.Consumers == nil {
+						cmd.val.Consumers = make(map[string]int64)
+					}
+					cmd.val.Consumers[consumerName] = consumerPending
+
+					return nil, nil
+				})
+				if err != nil {
+					return nil, err
+				}
+			}
+			return nil, nil
+		})
+		if err != nil && err != Nil {
+			return nil, err
+		}
+
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type XPendingExt struct {
+	ID         string
+	Consumer   string
+	Idle       time.Duration
+	RetryCount int64
+}
+
+type XPendingExtCmd struct {
+	baseCmd
+	val []XPendingExt
+}
+
+var _ Cmder = (*XPendingExtCmd)(nil)
+
+func NewXPendingExtCmd(ctx context.Context, args ...interface{}) *XPendingExtCmd {
+	return &XPendingExtCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *XPendingExtCmd) Val() []XPendingExt {
+	return cmd.val
+}
+
+func (cmd *XPendingExtCmd) Result() ([]XPendingExt, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XPendingExtCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XPendingExtCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]XPendingExt, 0, n)
+		for i := int64(0); i < n; i++ {
+			_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+				if n != 4 {
+					return nil, fmt.Errorf("got %d, wanted 4", n)
+				}
+
+				id, err := rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+
+				consumer, err := rd.ReadString()
+				if err != nil && err != Nil {
+					return nil, err
+				}
+
+				idle, err := rd.ReadIntReply()
+				if err != nil && err != Nil {
+					return nil, err
+				}
+
+				retryCount, err := rd.ReadIntReply()
+				if err != nil && err != Nil {
+					return nil, err
+				}
+
+				cmd.val = append(cmd.val, XPendingExt{
+					ID:         id,
+					Consumer:   consumer,
+					Idle:       time.Duration(idle) * time.Millisecond,
+					RetryCount: retryCount,
+				})
+				return nil, nil
+			})
+			if err != nil {
+				return nil, err
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type XInfoGroupsCmd struct {
+	baseCmd
+	val []XInfoGroup
+}
+
+type XInfoGroup struct {
+	Name            string
+	Consumers       int64
+	Pending         int64
+	LastDeliveredID string
+}
+
+var _ Cmder = (*XInfoGroupsCmd)(nil)
+
+func NewXInfoGroupsCmd(ctx context.Context, stream string) *XInfoGroupsCmd {
+	return &XInfoGroupsCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: []interface{}{"xinfo", "groups", stream},
+		},
+	}
+}
+
+func (cmd *XInfoGroupsCmd) Val() []XInfoGroup {
+	return cmd.val
+}
+
+func (cmd *XInfoGroupsCmd) Result() ([]XInfoGroup, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XInfoGroupsCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XInfoGroupsCmd) readReply(rd *proto.Reader) error {
+	n, err := rd.ReadArrayLen()
+	if err != nil {
+		return err
+	}
+
+	cmd.val = make([]XInfoGroup, n)
+
+	for i := 0; i < n; i++ {
+		cmd.val[i], err = readXGroupInfo(rd)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func readXGroupInfo(rd *proto.Reader) (XInfoGroup, error) {
+	var group XInfoGroup
+
+	n, err := rd.ReadArrayLen()
+	if err != nil {
+		return group, err
+	}
+	if n != 8 {
+		return group, fmt.Errorf("redis: got %d elements in XINFO GROUPS reply, wanted 8", n)
+	}
+
+	for i := 0; i < 4; i++ {
+		key, err := rd.ReadString()
+		if err != nil {
+			return group, err
+		}
+
+		val, err := rd.ReadString()
+		if err != nil {
+			return group, err
+		}
+
+		switch key {
+		case "name":
+			group.Name = val
+		case "consumers":
+			group.Consumers, err = strconv.ParseInt(val, 0, 64)
+			if err != nil {
+				return group, err
+			}
+		case "pending":
+			group.Pending, err = strconv.ParseInt(val, 0, 64)
+			if err != nil {
+				return group, err
+			}
+		case "last-delivered-id":
+			group.LastDeliveredID = val
+		default:
+			return group, fmt.Errorf("redis: unexpected content %s in XINFO GROUPS reply", key)
+		}
+	}
+
+	return group, nil
+}
+
+//------------------------------------------------------------------------------
+
+type XInfoStreamCmd struct {
+	baseCmd
+	val *XInfoStream
+}
+
+type XInfoStream struct {
+	Length          int64
+	RadixTreeKeys   int64
+	RadixTreeNodes  int64
+	Groups          int64
+	LastGeneratedID string
+	FirstEntry      XMessage
+	LastEntry       XMessage
+}
+
+var _ Cmder = (*XInfoStreamCmd)(nil)
+
+func NewXInfoStreamCmd(ctx context.Context, stream string) *XInfoStreamCmd {
+	return &XInfoStreamCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: []interface{}{"xinfo", "stream", stream},
+		},
+	}
+}
+
+func (cmd *XInfoStreamCmd) Val() *XInfoStream {
+	return cmd.val
+}
+
+func (cmd *XInfoStreamCmd) Result() (*XInfoStream, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *XInfoStreamCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *XInfoStreamCmd) readReply(rd *proto.Reader) error {
+	v, err := rd.ReadReply(xStreamInfoParser)
+	if err != nil {
+		return err
+	}
+	cmd.val = v.(*XInfoStream)
+	return nil
+}
+
+func xStreamInfoParser(rd *proto.Reader, n int64) (interface{}, error) {
+	if n != 14 {
+		return nil, fmt.Errorf("redis: got %d elements in XINFO STREAM reply,"+
+			"wanted 14", n)
+	}
+	var info XInfoStream
+	for i := 0; i < 7; i++ {
+		key, err := rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+		switch key {
+		case "length":
+			info.Length, err = rd.ReadIntReply()
+		case "radix-tree-keys":
+			info.RadixTreeKeys, err = rd.ReadIntReply()
+		case "radix-tree-nodes":
+			info.RadixTreeNodes, err = rd.ReadIntReply()
+		case "groups":
+			info.Groups, err = rd.ReadIntReply()
+		case "last-generated-id":
+			info.LastGeneratedID, err = rd.ReadString()
+		case "first-entry":
+			info.FirstEntry, err = readXMessage(rd)
+		case "last-entry":
+			info.LastEntry, err = readXMessage(rd)
+		default:
+			return nil, fmt.Errorf("redis: unexpected content %s "+
+				"in XINFO STREAM reply", key)
+		}
+		if err != nil {
+			return nil, err
+		}
+	}
+	return &info, nil
+}
+
+//------------------------------------------------------------------------------
+
+type ZSliceCmd struct {
+	baseCmd
+
+	val []Z
+}
+
+var _ Cmder = (*ZSliceCmd)(nil)
+
+func NewZSliceCmd(ctx context.Context, args ...interface{}) *ZSliceCmd {
+	return &ZSliceCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *ZSliceCmd) Val() []Z {
+	return cmd.val
+}
+
+func (cmd *ZSliceCmd) Result() ([]Z, error) {
+	return cmd.val, cmd.err
+}
+
+func (cmd *ZSliceCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *ZSliceCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]Z, n/2)
+		for i := 0; i < len(cmd.val); i++ {
+			member, err := rd.ReadString()
+			if err != nil {
+				return nil, err
+			}
+
+			score, err := rd.ReadFloatReply()
+			if err != nil {
+				return nil, err
+			}
+
+			cmd.val[i] = Z{
+				Member: member,
+				Score:  score,
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type ZWithKeyCmd struct {
+	baseCmd
+
+	val *ZWithKey
+}
+
+var _ Cmder = (*ZWithKeyCmd)(nil)
+
+func NewZWithKeyCmd(ctx context.Context, args ...interface{}) *ZWithKeyCmd {
+	return &ZWithKeyCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *ZWithKeyCmd) Val() *ZWithKey {
+	return cmd.val
+}
+
+func (cmd *ZWithKeyCmd) Result() (*ZWithKey, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *ZWithKeyCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *ZWithKeyCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		if n != 3 {
+			return nil, fmt.Errorf("got %d elements, expected 3", n)
+		}
+
+		cmd.val = &ZWithKey{}
+		var err error
+
+		cmd.val.Key, err = rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+
+		cmd.val.Member, err = rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+
+		cmd.val.Score, err = rd.ReadFloatReply()
+		if err != nil {
+			return nil, err
+		}
+
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type ScanCmd struct {
+	baseCmd
+
+	page   []string
+	cursor uint64
+
+	process cmdable
+}
+
+var _ Cmder = (*ScanCmd)(nil)
+
+func NewScanCmd(ctx context.Context, process cmdable, args ...interface{}) *ScanCmd {
+	return &ScanCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+		process: process,
+	}
+}
+
+func (cmd *ScanCmd) Val() (keys []string, cursor uint64) {
+	return cmd.page, cmd.cursor
+}
+
+func (cmd *ScanCmd) Result() (keys []string, cursor uint64, err error) {
+	return cmd.page, cmd.cursor, cmd.err
+}
+
+func (cmd *ScanCmd) String() string {
+	return cmdString(cmd, cmd.page)
+}
+
+func (cmd *ScanCmd) readReply(rd *proto.Reader) (err error) {
+	cmd.page, cmd.cursor, err = rd.ReadScanReply()
+	return err
+}
+
+// Iterator creates a new ScanIterator.
+func (cmd *ScanCmd) Iterator() *ScanIterator {
+	return &ScanIterator{
+		cmd: cmd,
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type ClusterNode struct {
+	ID   string
+	Addr string
+}
+
+type ClusterSlot struct {
+	Start int
+	End   int
+	Nodes []ClusterNode
+}
+
+type ClusterSlotsCmd struct {
+	baseCmd
+
+	val []ClusterSlot
+}
+
+var _ Cmder = (*ClusterSlotsCmd)(nil)
+
+func NewClusterSlotsCmd(ctx context.Context, args ...interface{}) *ClusterSlotsCmd {
+	return &ClusterSlotsCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *ClusterSlotsCmd) Val() []ClusterSlot {
+	return cmd.val
+}
+
+func (cmd *ClusterSlotsCmd) Result() ([]ClusterSlot, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *ClusterSlotsCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *ClusterSlotsCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]ClusterSlot, n)
+		for i := 0; i < len(cmd.val); i++ {
+			n, err := rd.ReadArrayLen()
+			if err != nil {
+				return nil, err
+			}
+			if n < 2 {
+				err := fmt.Errorf("redis: got %d elements in cluster info, expected at least 2", n)
+				return nil, err
+			}
+
+			start, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+
+			end, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+
+			nodes := make([]ClusterNode, n-2)
+			for j := 0; j < len(nodes); j++ {
+				n, err := rd.ReadArrayLen()
+				if err != nil {
+					return nil, err
+				}
+				if n != 2 && n != 3 {
+					err := fmt.Errorf("got %d elements in cluster info address, expected 2 or 3", n)
+					return nil, err
+				}
+
+				ip, err := rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+
+				port, err := rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+
+				nodes[j].Addr = net.JoinHostPort(ip, port)
+
+				if n == 3 {
+					id, err := rd.ReadString()
+					if err != nil {
+						return nil, err
+					}
+					nodes[j].ID = id
+				}
+			}
+
+			cmd.val[i] = ClusterSlot{
+				Start: int(start),
+				End:   int(end),
+				Nodes: nodes,
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+// GeoLocation is used with GeoAdd to add geospatial location.
+type GeoLocation struct {
+	Name                      string
+	Longitude, Latitude, Dist float64
+	GeoHash                   int64
+}
+
+// GeoRadiusQuery is used with GeoRadius to query geospatial index.
+type GeoRadiusQuery struct {
+	Radius float64
+	// Can be m, km, ft, or mi. Default is km.
+	Unit        string
+	WithCoord   bool
+	WithDist    bool
+	WithGeoHash bool
+	Count       int
+	// Can be ASC or DESC. Default is no sort order.
+	Sort      string
+	Store     string
+	StoreDist string
+}
+
+type GeoLocationCmd struct {
+	baseCmd
+
+	q         *GeoRadiusQuery
+	locations []GeoLocation
+}
+
+var _ Cmder = (*GeoLocationCmd)(nil)
+
+func NewGeoLocationCmd(ctx context.Context, q *GeoRadiusQuery, args ...interface{}) *GeoLocationCmd {
+	return &GeoLocationCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: geoLocationArgs(q, args...),
+		},
+		q: q,
+	}
+}
+
+func geoLocationArgs(q *GeoRadiusQuery, args ...interface{}) []interface{} {
+	args = append(args, q.Radius)
+	if q.Unit != "" {
+		args = append(args, q.Unit)
+	} else {
+		args = append(args, "km")
+	}
+	if q.WithCoord {
+		args = append(args, "withcoord")
+	}
+	if q.WithDist {
+		args = append(args, "withdist")
+	}
+	if q.WithGeoHash {
+		args = append(args, "withhash")
+	}
+	if q.Count > 0 {
+		args = append(args, "count", q.Count)
+	}
+	if q.Sort != "" {
+		args = append(args, q.Sort)
+	}
+	if q.Store != "" {
+		args = append(args, "store")
+		args = append(args, q.Store)
+	}
+	if q.StoreDist != "" {
+		args = append(args, "storedist")
+		args = append(args, q.StoreDist)
+	}
+	return args
+}
+
+func (cmd *GeoLocationCmd) Val() []GeoLocation {
+	return cmd.locations
+}
+
+func (cmd *GeoLocationCmd) Result() ([]GeoLocation, error) {
+	return cmd.locations, cmd.err
+}
+
+func (cmd *GeoLocationCmd) String() string {
+	return cmdString(cmd, cmd.locations)
+}
+
+func (cmd *GeoLocationCmd) readReply(rd *proto.Reader) error {
+	v, err := rd.ReadArrayReply(newGeoLocationSliceParser(cmd.q))
+	if err != nil {
+		return err
+	}
+	cmd.locations = v.([]GeoLocation)
+	return nil
+}
+
+func newGeoLocationSliceParser(q *GeoRadiusQuery) proto.MultiBulkParse {
+	return func(rd *proto.Reader, n int64) (interface{}, error) {
+		locs := make([]GeoLocation, 0, n)
+		for i := int64(0); i < n; i++ {
+			v, err := rd.ReadReply(newGeoLocationParser(q))
+			if err != nil {
+				return nil, err
+			}
+			switch vv := v.(type) {
+			case string:
+				locs = append(locs, GeoLocation{
+					Name: vv,
+				})
+			case *GeoLocation:
+				// TODO: avoid copying
+				locs = append(locs, *vv)
+			default:
+				return nil, fmt.Errorf("got %T, expected string or *GeoLocation", v)
+			}
+		}
+		return locs, nil
+	}
+}
+
+func newGeoLocationParser(q *GeoRadiusQuery) proto.MultiBulkParse {
+	return func(rd *proto.Reader, n int64) (interface{}, error) {
+		var loc GeoLocation
+		var err error
+
+		loc.Name, err = rd.ReadString()
+		if err != nil {
+			return nil, err
+		}
+		if q.WithDist {
+			loc.Dist, err = rd.ReadFloatReply()
+			if err != nil {
+				return nil, err
+			}
+		}
+		if q.WithGeoHash {
+			loc.GeoHash, err = rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+		}
+		if q.WithCoord {
+			n, err := rd.ReadArrayLen()
+			if err != nil {
+				return nil, err
+			}
+			if n != 2 {
+				return nil, fmt.Errorf("got %d coordinates, expected 2", n)
+			}
+
+			loc.Longitude, err = rd.ReadFloatReply()
+			if err != nil {
+				return nil, err
+			}
+			loc.Latitude, err = rd.ReadFloatReply()
+			if err != nil {
+				return nil, err
+			}
+		}
+
+		return &loc, nil
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type GeoPos struct {
+	Longitude, Latitude float64
+}
+
+type GeoPosCmd struct {
+	baseCmd
+
+	val []*GeoPos
+}
+
+var _ Cmder = (*GeoPosCmd)(nil)
+
+func NewGeoPosCmd(ctx context.Context, args ...interface{}) *GeoPosCmd {
+	return &GeoPosCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *GeoPosCmd) Val() []*GeoPos {
+	return cmd.val
+}
+
+func (cmd *GeoPosCmd) Result() ([]*GeoPos, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *GeoPosCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *GeoPosCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]*GeoPos, n)
+		for i := 0; i < len(cmd.val); i++ {
+			i := i
+			_, err := rd.ReadReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+				longitude, err := rd.ReadFloatReply()
+				if err != nil {
+					return nil, err
+				}
+
+				latitude, err := rd.ReadFloatReply()
+				if err != nil {
+					return nil, err
+				}
+
+				cmd.val[i] = &GeoPos{
+					Longitude: longitude,
+					Latitude:  latitude,
+				}
+				return nil, nil
+			})
+			if err != nil {
+				if err == Nil {
+					cmd.val[i] = nil
+					continue
+				}
+				return nil, err
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
+
+//------------------------------------------------------------------------------
+
+type CommandInfo struct {
+	Name        string
+	Arity       int8
+	Flags       []string
+	ACLFlags    []string
+	FirstKeyPos int8
+	LastKeyPos  int8
+	StepCount   int8
+	ReadOnly    bool
+}
+
+type CommandsInfoCmd struct {
+	baseCmd
+
+	val map[string]*CommandInfo
+}
+
+var _ Cmder = (*CommandsInfoCmd)(nil)
+
+func NewCommandsInfoCmd(ctx context.Context, args ...interface{}) *CommandsInfoCmd {
+	return &CommandsInfoCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *CommandsInfoCmd) Val() map[string]*CommandInfo {
+	return cmd.val
+}
+
+func (cmd *CommandsInfoCmd) Result() (map[string]*CommandInfo, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *CommandsInfoCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *CommandsInfoCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make(map[string]*CommandInfo, n)
+		for i := int64(0); i < n; i++ {
+			v, err := rd.ReadReply(commandInfoParser)
+			if err != nil {
+				return nil, err
+			}
+			vv := v.(*CommandInfo)
+			cmd.val[vv.Name] = vv
+		}
+		return nil, nil
+	})
+	return err
+}
+
+func commandInfoParser(rd *proto.Reader, n int64) (interface{}, error) {
+	const numArgRedis5 = 6
+	const numArgRedis6 = 7
+
+	switch n {
+	case numArgRedis5, numArgRedis6:
+		// continue
+	default:
+		return nil, fmt.Errorf("redis: got %d elements in COMMAND reply, wanted 7", n)
+	}
+
+	var cmd CommandInfo
+	var err error
+
+	cmd.Name, err = rd.ReadString()
+	if err != nil {
+		return nil, err
+	}
+
+	arity, err := rd.ReadIntReply()
+	if err != nil {
+		return nil, err
+	}
+	cmd.Arity = int8(arity)
+
+	_, err = rd.ReadReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.Flags = make([]string, n)
+		for i := 0; i < len(cmd.Flags); i++ {
+			switch s, err := rd.ReadString(); {
+			case err == Nil:
+				cmd.Flags[i] = ""
+			case err != nil:
+				return nil, err
+			default:
+				cmd.Flags[i] = s
+			}
+		}
+		return nil, nil
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	firstKeyPos, err := rd.ReadIntReply()
+	if err != nil {
+		return nil, err
+	}
+	cmd.FirstKeyPos = int8(firstKeyPos)
+
+	lastKeyPos, err := rd.ReadIntReply()
+	if err != nil {
+		return nil, err
+	}
+	cmd.LastKeyPos = int8(lastKeyPos)
+
+	stepCount, err := rd.ReadIntReply()
+	if err != nil {
+		return nil, err
+	}
+	cmd.StepCount = int8(stepCount)
+
+	for _, flag := range cmd.Flags {
+		if flag == "readonly" {
+			cmd.ReadOnly = true
+			break
+		}
+	}
+
+	if n == numArgRedis5 {
+		return &cmd, nil
+	}
+
+	_, err = rd.ReadReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.ACLFlags = make([]string, n)
+		for i := 0; i < len(cmd.ACLFlags); i++ {
+			switch s, err := rd.ReadString(); {
+			case err == Nil:
+				cmd.ACLFlags[i] = ""
+			case err != nil:
+				return nil, err
+			default:
+				cmd.ACLFlags[i] = s
+			}
+		}
+		return nil, nil
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	return &cmd, nil
+}
+
+//------------------------------------------------------------------------------
+
+type cmdsInfoCache struct {
+	fn func(ctx context.Context) (map[string]*CommandInfo, error)
+
+	once internal.Once
+	cmds map[string]*CommandInfo
+}
+
+func newCmdsInfoCache(fn func(ctx context.Context) (map[string]*CommandInfo, error)) *cmdsInfoCache {
+	return &cmdsInfoCache{
+		fn: fn,
+	}
+}
+
+func (c *cmdsInfoCache) Get(ctx context.Context) (map[string]*CommandInfo, error) {
+	err := c.once.Do(func() error {
+		cmds, err := c.fn(ctx)
+		if err != nil {
+			return err
+		}
+
+		// Extensions have cmd names in upper case. Convert them to lower case.
+		for k, v := range cmds {
+			lower := internal.ToLower(k)
+			if lower != k {
+				cmds[lower] = v
+			}
+		}
+
+		c.cmds = cmds
+		return nil
+	})
+	return c.cmds, err
+}
+
+//------------------------------------------------------------------------------
+
+type SlowLog struct {
+	ID       int64
+	Time     time.Time
+	Duration time.Duration
+	Args     []string
+	// These are also optional fields emitted only by Redis 4.0 or greater:
+	// https://redis.io/commands/slowlog#output-format
+	ClientAddr string
+	ClientName string
+}
+
+type SlowLogCmd struct {
+	baseCmd
+
+	val []SlowLog
+}
+
+var _ Cmder = (*SlowLogCmd)(nil)
+
+func NewSlowLogCmd(ctx context.Context, args ...interface{}) *SlowLogCmd {
+	return &SlowLogCmd{
+		baseCmd: baseCmd{
+			ctx:  ctx,
+			args: args,
+		},
+	}
+}
+
+func (cmd *SlowLogCmd) Val() []SlowLog {
+	return cmd.val
+}
+
+func (cmd *SlowLogCmd) Result() ([]SlowLog, error) {
+	return cmd.Val(), cmd.Err()
+}
+
+func (cmd *SlowLogCmd) String() string {
+	return cmdString(cmd, cmd.val)
+}
+
+func (cmd *SlowLogCmd) readReply(rd *proto.Reader) error {
+	_, err := rd.ReadArrayReply(func(rd *proto.Reader, n int64) (interface{}, error) {
+		cmd.val = make([]SlowLog, n)
+		for i := 0; i < len(cmd.val); i++ {
+			n, err := rd.ReadArrayLen()
+			if err != nil {
+				return nil, err
+			}
+			if n < 4 {
+				err := fmt.Errorf("redis: got %d elements in slowlog get, expected at least 4", n)
+				return nil, err
+			}
+
+			id, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+
+			createdAt, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+			createdAtTime := time.Unix(createdAt, 0)
+
+			costs, err := rd.ReadIntReply()
+			if err != nil {
+				return nil, err
+			}
+			costsDuration := time.Duration(costs) * time.Microsecond
+
+			cmdLen, err := rd.ReadArrayLen()
+			if err != nil {
+				return nil, err
+			}
+			if cmdLen < 1 {
+				err := fmt.Errorf("redis: got %d elements commands reply in slowlog get, expected at least 1", cmdLen)
+				return nil, err
+			}
+
+			cmdString := make([]string, cmdLen)
+			for i := 0; i < cmdLen; i++ {
+				cmdString[i], err = rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+			}
+
+			var address, name string
+			for i := 4; i < n; i++ {
+				str, err := rd.ReadString()
+				if err != nil {
+					return nil, err
+				}
+				if i == 4 {
+					address = str
+				} else if i == 5 {
+					name = str
+				}
+			}
+
+			cmd.val[i] = SlowLog{
+				ID:         id,
+				Time:       createdAtTime,
+				Duration:   costsDuration,
+				Args:       cmdString,
+				ClientAddr: address,
+				ClientName: name,
+			}
+		}
+		return nil, nil
+	})
+	return err
+}
diff --git a/vendor/github.com/go-redis/redis/v8/commands.go b/vendor/github.com/go-redis/redis/v8/commands.go
new file mode 100644
index 0000000..79698ba
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/commands.go
@@ -0,0 +1,2773 @@
+package redis
+
+import (
+	"context"
+	"errors"
+	"io"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+)
+
+// KeepTTL is an option for Set command to keep key's existing TTL.
+// For example:
+//
+//    rdb.Set(ctx, key, value, redis.KeepTTL)
+const KeepTTL = -1
+
+func usePrecise(dur time.Duration) bool {
+	return dur < time.Second || dur%time.Second != 0
+}
+
+func formatMs(ctx context.Context, dur time.Duration) int64 {
+	if dur > 0 && dur < time.Millisecond {
+		internal.Logger.Printf(
+			ctx,
+			"specified duration is %s, but minimal supported value is %s - truncating to 1ms",
+			dur, time.Millisecond,
+		)
+		return 1
+	}
+	return int64(dur / time.Millisecond)
+}
+
+func formatSec(ctx context.Context, dur time.Duration) int64 {
+	if dur > 0 && dur < time.Second {
+		internal.Logger.Printf(
+			ctx,
+			"specified duration is %s, but minimal supported value is %s - truncating to 1s",
+			dur, time.Second,
+		)
+		return 1
+	}
+	return int64(dur / time.Second)
+}
+
+func appendArgs(dst, src []interface{}) []interface{} {
+	if len(src) == 1 {
+		return appendArg(dst, src[0])
+	}
+
+	dst = append(dst, src...)
+	return dst
+}
+
+func appendArg(dst []interface{}, arg interface{}) []interface{} {
+	switch arg := arg.(type) {
+	case []string:
+		for _, s := range arg {
+			dst = append(dst, s)
+		}
+		return dst
+	case []interface{}:
+		dst = append(dst, arg...)
+		return dst
+	case map[string]interface{}:
+		for k, v := range arg {
+			dst = append(dst, k, v)
+		}
+		return dst
+	default:
+		return append(dst, arg)
+	}
+}
+
+type Cmdable interface {
+	Pipeline() Pipeliner
+	Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error)
+
+	TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error)
+	TxPipeline() Pipeliner
+
+	Command(ctx context.Context) *CommandsInfoCmd
+	ClientGetName(ctx context.Context) *StringCmd
+	Echo(ctx context.Context, message interface{}) *StringCmd
+	Ping(ctx context.Context) *StatusCmd
+	Quit(ctx context.Context) *StatusCmd
+	Del(ctx context.Context, keys ...string) *IntCmd
+	Unlink(ctx context.Context, keys ...string) *IntCmd
+	Dump(ctx context.Context, key string) *StringCmd
+	Exists(ctx context.Context, keys ...string) *IntCmd
+	Expire(ctx context.Context, key string, expiration time.Duration) *BoolCmd
+	ExpireAt(ctx context.Context, key string, tm time.Time) *BoolCmd
+	Keys(ctx context.Context, pattern string) *StringSliceCmd
+	Migrate(ctx context.Context, host, port, key string, db int, timeout time.Duration) *StatusCmd
+	Move(ctx context.Context, key string, db int) *BoolCmd
+	ObjectRefCount(ctx context.Context, key string) *IntCmd
+	ObjectEncoding(ctx context.Context, key string) *StringCmd
+	ObjectIdleTime(ctx context.Context, key string) *DurationCmd
+	Persist(ctx context.Context, key string) *BoolCmd
+	PExpire(ctx context.Context, key string, expiration time.Duration) *BoolCmd
+	PExpireAt(ctx context.Context, key string, tm time.Time) *BoolCmd
+	PTTL(ctx context.Context, key string) *DurationCmd
+	RandomKey(ctx context.Context) *StringCmd
+	Rename(ctx context.Context, key, newkey string) *StatusCmd
+	RenameNX(ctx context.Context, key, newkey string) *BoolCmd
+	Restore(ctx context.Context, key string, ttl time.Duration, value string) *StatusCmd
+	RestoreReplace(ctx context.Context, key string, ttl time.Duration, value string) *StatusCmd
+	Sort(ctx context.Context, key string, sort *Sort) *StringSliceCmd
+	SortStore(ctx context.Context, key, store string, sort *Sort) *IntCmd
+	SortInterfaces(ctx context.Context, key string, sort *Sort) *SliceCmd
+	Touch(ctx context.Context, keys ...string) *IntCmd
+	TTL(ctx context.Context, key string) *DurationCmd
+	Type(ctx context.Context, key string) *StatusCmd
+	Append(ctx context.Context, key, value string) *IntCmd
+	Decr(ctx context.Context, key string) *IntCmd
+	DecrBy(ctx context.Context, key string, decrement int64) *IntCmd
+	Get(ctx context.Context, key string) *StringCmd
+	GetRange(ctx context.Context, key string, start, end int64) *StringCmd
+	GetSet(ctx context.Context, key string, value interface{}) *StringCmd
+	Incr(ctx context.Context, key string) *IntCmd
+	IncrBy(ctx context.Context, key string, value int64) *IntCmd
+	IncrByFloat(ctx context.Context, key string, value float64) *FloatCmd
+	MGet(ctx context.Context, keys ...string) *SliceCmd
+	MSet(ctx context.Context, values ...interface{}) *StatusCmd
+	MSetNX(ctx context.Context, values ...interface{}) *BoolCmd
+	Set(ctx context.Context, key string, value interface{}, expiration time.Duration) *StatusCmd
+	SetEX(ctx context.Context, key string, value interface{}, expiration time.Duration) *StatusCmd
+	SetNX(ctx context.Context, key string, value interface{}, expiration time.Duration) *BoolCmd
+	SetXX(ctx context.Context, key string, value interface{}, expiration time.Duration) *BoolCmd
+	SetRange(ctx context.Context, key string, offset int64, value string) *IntCmd
+	StrLen(ctx context.Context, key string) *IntCmd
+
+	GetBit(ctx context.Context, key string, offset int64) *IntCmd
+	SetBit(ctx context.Context, key string, offset int64, value int) *IntCmd
+	BitCount(ctx context.Context, key string, bitCount *BitCount) *IntCmd
+	BitOpAnd(ctx context.Context, destKey string, keys ...string) *IntCmd
+	BitOpOr(ctx context.Context, destKey string, keys ...string) *IntCmd
+	BitOpXor(ctx context.Context, destKey string, keys ...string) *IntCmd
+	BitOpNot(ctx context.Context, destKey string, key string) *IntCmd
+	BitPos(ctx context.Context, key string, bit int64, pos ...int64) *IntCmd
+	BitField(ctx context.Context, key string, args ...interface{}) *IntSliceCmd
+
+	Scan(ctx context.Context, cursor uint64, match string, count int64) *ScanCmd
+	SScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd
+	HScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd
+	ZScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd
+
+	HDel(ctx context.Context, key string, fields ...string) *IntCmd
+	HExists(ctx context.Context, key, field string) *BoolCmd
+	HGet(ctx context.Context, key, field string) *StringCmd
+	HGetAll(ctx context.Context, key string) *StringStringMapCmd
+	HIncrBy(ctx context.Context, key, field string, incr int64) *IntCmd
+	HIncrByFloat(ctx context.Context, key, field string, incr float64) *FloatCmd
+	HKeys(ctx context.Context, key string) *StringSliceCmd
+	HLen(ctx context.Context, key string) *IntCmd
+	HMGet(ctx context.Context, key string, fields ...string) *SliceCmd
+	HSet(ctx context.Context, key string, values ...interface{}) *IntCmd
+	HMSet(ctx context.Context, key string, values ...interface{}) *BoolCmd
+	HSetNX(ctx context.Context, key, field string, value interface{}) *BoolCmd
+	HVals(ctx context.Context, key string) *StringSliceCmd
+
+	BLPop(ctx context.Context, timeout time.Duration, keys ...string) *StringSliceCmd
+	BRPop(ctx context.Context, timeout time.Duration, keys ...string) *StringSliceCmd
+	BRPopLPush(ctx context.Context, source, destination string, timeout time.Duration) *StringCmd
+	LIndex(ctx context.Context, key string, index int64) *StringCmd
+	LInsert(ctx context.Context, key, op string, pivot, value interface{}) *IntCmd
+	LInsertBefore(ctx context.Context, key string, pivot, value interface{}) *IntCmd
+	LInsertAfter(ctx context.Context, key string, pivot, value interface{}) *IntCmd
+	LLen(ctx context.Context, key string) *IntCmd
+	LPop(ctx context.Context, key string) *StringCmd
+	LPos(ctx context.Context, key string, value string, args LPosArgs) *IntCmd
+	LPosCount(ctx context.Context, key string, value string, count int64, args LPosArgs) *IntSliceCmd
+	LPush(ctx context.Context, key string, values ...interface{}) *IntCmd
+	LPushX(ctx context.Context, key string, values ...interface{}) *IntCmd
+	LRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd
+	LRem(ctx context.Context, key string, count int64, value interface{}) *IntCmd
+	LSet(ctx context.Context, key string, index int64, value interface{}) *StatusCmd
+	LTrim(ctx context.Context, key string, start, stop int64) *StatusCmd
+	RPop(ctx context.Context, key string) *StringCmd
+	RPopLPush(ctx context.Context, source, destination string) *StringCmd
+	RPush(ctx context.Context, key string, values ...interface{}) *IntCmd
+	RPushX(ctx context.Context, key string, values ...interface{}) *IntCmd
+
+	SAdd(ctx context.Context, key string, members ...interface{}) *IntCmd
+	SCard(ctx context.Context, key string) *IntCmd
+	SDiff(ctx context.Context, keys ...string) *StringSliceCmd
+	SDiffStore(ctx context.Context, destination string, keys ...string) *IntCmd
+	SInter(ctx context.Context, keys ...string) *StringSliceCmd
+	SInterStore(ctx context.Context, destination string, keys ...string) *IntCmd
+	SIsMember(ctx context.Context, key string, member interface{}) *BoolCmd
+	SMembers(ctx context.Context, key string) *StringSliceCmd
+	SMembersMap(ctx context.Context, key string) *StringStructMapCmd
+	SMove(ctx context.Context, source, destination string, member interface{}) *BoolCmd
+	SPop(ctx context.Context, key string) *StringCmd
+	SPopN(ctx context.Context, key string, count int64) *StringSliceCmd
+	SRandMember(ctx context.Context, key string) *StringCmd
+	SRandMemberN(ctx context.Context, key string, count int64) *StringSliceCmd
+	SRem(ctx context.Context, key string, members ...interface{}) *IntCmd
+	SUnion(ctx context.Context, keys ...string) *StringSliceCmd
+	SUnionStore(ctx context.Context, destination string, keys ...string) *IntCmd
+
+	XAdd(ctx context.Context, a *XAddArgs) *StringCmd
+	XDel(ctx context.Context, stream string, ids ...string) *IntCmd
+	XLen(ctx context.Context, stream string) *IntCmd
+	XRange(ctx context.Context, stream, start, stop string) *XMessageSliceCmd
+	XRangeN(ctx context.Context, stream, start, stop string, count int64) *XMessageSliceCmd
+	XRevRange(ctx context.Context, stream string, start, stop string) *XMessageSliceCmd
+	XRevRangeN(ctx context.Context, stream string, start, stop string, count int64) *XMessageSliceCmd
+	XRead(ctx context.Context, a *XReadArgs) *XStreamSliceCmd
+	XReadStreams(ctx context.Context, streams ...string) *XStreamSliceCmd
+	XGroupCreate(ctx context.Context, stream, group, start string) *StatusCmd
+	XGroupCreateMkStream(ctx context.Context, stream, group, start string) *StatusCmd
+	XGroupSetID(ctx context.Context, stream, group, start string) *StatusCmd
+	XGroupDestroy(ctx context.Context, stream, group string) *IntCmd
+	XGroupDelConsumer(ctx context.Context, stream, group, consumer string) *IntCmd
+	XReadGroup(ctx context.Context, a *XReadGroupArgs) *XStreamSliceCmd
+	XAck(ctx context.Context, stream, group string, ids ...string) *IntCmd
+	XPending(ctx context.Context, stream, group string) *XPendingCmd
+	XPendingExt(ctx context.Context, a *XPendingExtArgs) *XPendingExtCmd
+	XClaim(ctx context.Context, a *XClaimArgs) *XMessageSliceCmd
+	XClaimJustID(ctx context.Context, a *XClaimArgs) *StringSliceCmd
+	XTrim(ctx context.Context, key string, maxLen int64) *IntCmd
+	XTrimApprox(ctx context.Context, key string, maxLen int64) *IntCmd
+	XInfoGroups(ctx context.Context, key string) *XInfoGroupsCmd
+	XInfoStream(ctx context.Context, key string) *XInfoStreamCmd
+
+	BZPopMax(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd
+	BZPopMin(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd
+	ZAdd(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZAddNX(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZAddXX(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZAddCh(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZAddNXCh(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZAddXXCh(ctx context.Context, key string, members ...*Z) *IntCmd
+	ZIncr(ctx context.Context, key string, member *Z) *FloatCmd
+	ZIncrNX(ctx context.Context, key string, member *Z) *FloatCmd
+	ZIncrXX(ctx context.Context, key string, member *Z) *FloatCmd
+	ZCard(ctx context.Context, key string) *IntCmd
+	ZCount(ctx context.Context, key, min, max string) *IntCmd
+	ZLexCount(ctx context.Context, key, min, max string) *IntCmd
+	ZIncrBy(ctx context.Context, key string, increment float64, member string) *FloatCmd
+	ZInterStore(ctx context.Context, destination string, store *ZStore) *IntCmd
+	ZPopMax(ctx context.Context, key string, count ...int64) *ZSliceCmd
+	ZPopMin(ctx context.Context, key string, count ...int64) *ZSliceCmd
+	ZRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd
+	ZRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd
+	ZRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
+	ZRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
+	ZRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd
+	ZRank(ctx context.Context, key, member string) *IntCmd
+	ZRem(ctx context.Context, key string, members ...interface{}) *IntCmd
+	ZRemRangeByRank(ctx context.Context, key string, start, stop int64) *IntCmd
+	ZRemRangeByScore(ctx context.Context, key, min, max string) *IntCmd
+	ZRemRangeByLex(ctx context.Context, key, min, max string) *IntCmd
+	ZRevRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd
+	ZRevRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd
+	ZRevRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
+	ZRevRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
+	ZRevRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd
+	ZRevRank(ctx context.Context, key, member string) *IntCmd
+	ZScore(ctx context.Context, key, member string) *FloatCmd
+	ZUnionStore(ctx context.Context, dest string, store *ZStore) *IntCmd
+
+	PFAdd(ctx context.Context, key string, els ...interface{}) *IntCmd
+	PFCount(ctx context.Context, keys ...string) *IntCmd
+	PFMerge(ctx context.Context, dest string, keys ...string) *StatusCmd
+
+	BgRewriteAOF(ctx context.Context) *StatusCmd
+	BgSave(ctx context.Context) *StatusCmd
+	ClientKill(ctx context.Context, ipPort string) *StatusCmd
+	ClientKillByFilter(ctx context.Context, keys ...string) *IntCmd
+	ClientList(ctx context.Context) *StringCmd
+	ClientPause(ctx context.Context, dur time.Duration) *BoolCmd
+	ClientID(ctx context.Context) *IntCmd
+	ConfigGet(ctx context.Context, parameter string) *SliceCmd
+	ConfigResetStat(ctx context.Context) *StatusCmd
+	ConfigSet(ctx context.Context, parameter, value string) *StatusCmd
+	ConfigRewrite(ctx context.Context) *StatusCmd
+	DBSize(ctx context.Context) *IntCmd
+	FlushAll(ctx context.Context) *StatusCmd
+	FlushAllAsync(ctx context.Context) *StatusCmd
+	FlushDB(ctx context.Context) *StatusCmd
+	FlushDBAsync(ctx context.Context) *StatusCmd
+	Info(ctx context.Context, section ...string) *StringCmd
+	LastSave(ctx context.Context) *IntCmd
+	Save(ctx context.Context) *StatusCmd
+	Shutdown(ctx context.Context) *StatusCmd
+	ShutdownSave(ctx context.Context) *StatusCmd
+	ShutdownNoSave(ctx context.Context) *StatusCmd
+	SlaveOf(ctx context.Context, host, port string) *StatusCmd
+	Time(ctx context.Context) *TimeCmd
+	DebugObject(ctx context.Context, key string) *StringCmd
+	ReadOnly(ctx context.Context) *StatusCmd
+	ReadWrite(ctx context.Context) *StatusCmd
+	MemoryUsage(ctx context.Context, key string, samples ...int) *IntCmd
+
+	Eval(ctx context.Context, script string, keys []string, args ...interface{}) *Cmd
+	EvalSha(ctx context.Context, sha1 string, keys []string, args ...interface{}) *Cmd
+	ScriptExists(ctx context.Context, hashes ...string) *BoolSliceCmd
+	ScriptFlush(ctx context.Context) *StatusCmd
+	ScriptKill(ctx context.Context) *StatusCmd
+	ScriptLoad(ctx context.Context, script string) *StringCmd
+
+	Publish(ctx context.Context, channel string, message interface{}) *IntCmd
+	PubSubChannels(ctx context.Context, pattern string) *StringSliceCmd
+	PubSubNumSub(ctx context.Context, channels ...string) *StringIntMapCmd
+	PubSubNumPat(ctx context.Context) *IntCmd
+
+	ClusterSlots(ctx context.Context) *ClusterSlotsCmd
+	ClusterNodes(ctx context.Context) *StringCmd
+	ClusterMeet(ctx context.Context, host, port string) *StatusCmd
+	ClusterForget(ctx context.Context, nodeID string) *StatusCmd
+	ClusterReplicate(ctx context.Context, nodeID string) *StatusCmd
+	ClusterResetSoft(ctx context.Context) *StatusCmd
+	ClusterResetHard(ctx context.Context) *StatusCmd
+	ClusterInfo(ctx context.Context) *StringCmd
+	ClusterKeySlot(ctx context.Context, key string) *IntCmd
+	ClusterGetKeysInSlot(ctx context.Context, slot int, count int) *StringSliceCmd
+	ClusterCountFailureReports(ctx context.Context, nodeID string) *IntCmd
+	ClusterCountKeysInSlot(ctx context.Context, slot int) *IntCmd
+	ClusterDelSlots(ctx context.Context, slots ...int) *StatusCmd
+	ClusterDelSlotsRange(ctx context.Context, min, max int) *StatusCmd
+	ClusterSaveConfig(ctx context.Context) *StatusCmd
+	ClusterSlaves(ctx context.Context, nodeID string) *StringSliceCmd
+	ClusterFailover(ctx context.Context) *StatusCmd
+	ClusterAddSlots(ctx context.Context, slots ...int) *StatusCmd
+	ClusterAddSlotsRange(ctx context.Context, min, max int) *StatusCmd
+
+	GeoAdd(ctx context.Context, key string, geoLocation ...*GeoLocation) *IntCmd
+	GeoPos(ctx context.Context, key string, members ...string) *GeoPosCmd
+	GeoRadius(ctx context.Context, key string, longitude, latitude float64, query *GeoRadiusQuery) *GeoLocationCmd
+	GeoRadiusStore(ctx context.Context, key string, longitude, latitude float64, query *GeoRadiusQuery) *IntCmd
+	GeoRadiusByMember(ctx context.Context, key, member string, query *GeoRadiusQuery) *GeoLocationCmd
+	GeoRadiusByMemberStore(ctx context.Context, key, member string, query *GeoRadiusQuery) *IntCmd
+	GeoDist(ctx context.Context, key string, member1, member2, unit string) *FloatCmd
+	GeoHash(ctx context.Context, key string, members ...string) *StringSliceCmd
+}
+
+type StatefulCmdable interface {
+	Cmdable
+	Auth(ctx context.Context, password string) *StatusCmd
+	AuthACL(ctx context.Context, username, password string) *StatusCmd
+	Select(ctx context.Context, index int) *StatusCmd
+	SwapDB(ctx context.Context, index1, index2 int) *StatusCmd
+	ClientSetName(ctx context.Context, name string) *BoolCmd
+}
+
+var (
+	_ Cmdable = (*Client)(nil)
+	_ Cmdable = (*Tx)(nil)
+	_ Cmdable = (*Ring)(nil)
+	_ Cmdable = (*ClusterClient)(nil)
+)
+
+type cmdable func(ctx context.Context, cmd Cmder) error
+
+type statefulCmdable func(ctx context.Context, cmd Cmder) error
+
+//------------------------------------------------------------------------------
+
+func (c statefulCmdable) Auth(ctx context.Context, password string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "auth", password)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Perform an AUTH command, using the given user and pass.
+// Should be used to authenticate the current connection with one of the connections defined in the ACL list
+// when connecting to a Redis 6.0 instance, or greater, that is using the Redis ACL system.
+func (c statefulCmdable) AuthACL(ctx context.Context, username, password string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "auth", username, password)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Wait(ctx context.Context, numSlaves int, timeout time.Duration) *IntCmd {
+	cmd := NewIntCmd(ctx, "wait", numSlaves, int(timeout/time.Millisecond))
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c statefulCmdable) Select(ctx context.Context, index int) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "select", index)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c statefulCmdable) SwapDB(ctx context.Context, index1, index2 int) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "swapdb", index1, index2)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// ClientSetName assigns a name to the connection.
+func (c statefulCmdable) ClientSetName(ctx context.Context, name string) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "client", "setname", name)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) Command(ctx context.Context) *CommandsInfoCmd {
+	cmd := NewCommandsInfoCmd(ctx, "command")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// ClientGetName returns the name of the connection.
+func (c cmdable) ClientGetName(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "client", "getname")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Echo(ctx context.Context, message interface{}) *StringCmd {
+	cmd := NewStringCmd(ctx, "echo", message)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Ping(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "ping")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Quit(ctx context.Context) *StatusCmd {
+	panic("not implemented")
+}
+
+func (c cmdable) Del(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "del"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Unlink(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "unlink"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Dump(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "dump", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Exists(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "exists"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Expire(ctx context.Context, key string, expiration time.Duration) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "expire", key, formatSec(ctx, expiration))
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ExpireAt(ctx context.Context, key string, tm time.Time) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "expireat", key, tm.Unix())
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Keys(ctx context.Context, pattern string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "keys", pattern)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Migrate(ctx context.Context, host, port, key string, db int, timeout time.Duration) *StatusCmd {
+	cmd := NewStatusCmd(
+		ctx,
+		"migrate",
+		host,
+		port,
+		key,
+		db,
+		formatMs(ctx, timeout),
+	)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Move(ctx context.Context, key string, db int) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "move", key, db)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ObjectRefCount(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "object", "refcount", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ObjectEncoding(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "object", "encoding", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ObjectIdleTime(ctx context.Context, key string) *DurationCmd {
+	cmd := NewDurationCmd(ctx, time.Second, "object", "idletime", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Persist(ctx context.Context, key string) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "persist", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PExpire(ctx context.Context, key string, expiration time.Duration) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "pexpire", key, formatMs(ctx, expiration))
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PExpireAt(ctx context.Context, key string, tm time.Time) *BoolCmd {
+	cmd := NewBoolCmd(
+		ctx,
+		"pexpireat",
+		key,
+		tm.UnixNano()/int64(time.Millisecond),
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PTTL(ctx context.Context, key string) *DurationCmd {
+	cmd := NewDurationCmd(ctx, time.Millisecond, "pttl", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RandomKey(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "randomkey")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Rename(ctx context.Context, key, newkey string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "rename", key, newkey)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RenameNX(ctx context.Context, key, newkey string) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "renamenx", key, newkey)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Restore(ctx context.Context, key string, ttl time.Duration, value string) *StatusCmd {
+	cmd := NewStatusCmd(
+		ctx,
+		"restore",
+		key,
+		formatMs(ctx, ttl),
+		value,
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RestoreReplace(ctx context.Context, key string, ttl time.Duration, value string) *StatusCmd {
+	cmd := NewStatusCmd(
+		ctx,
+		"restore",
+		key,
+		formatMs(ctx, ttl),
+		value,
+		"replace",
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type Sort struct {
+	By            string
+	Offset, Count int64
+	Get           []string
+	Order         string
+	Alpha         bool
+}
+
+func (sort *Sort) args(key string) []interface{} {
+	args := []interface{}{"sort", key}
+	if sort.By != "" {
+		args = append(args, "by", sort.By)
+	}
+	if sort.Offset != 0 || sort.Count != 0 {
+		args = append(args, "limit", sort.Offset, sort.Count)
+	}
+	for _, get := range sort.Get {
+		args = append(args, "get", get)
+	}
+	if sort.Order != "" {
+		args = append(args, sort.Order)
+	}
+	if sort.Alpha {
+		args = append(args, "alpha")
+	}
+	return args
+}
+
+func (c cmdable) Sort(ctx context.Context, key string, sort *Sort) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, sort.args(key)...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SortStore(ctx context.Context, key, store string, sort *Sort) *IntCmd {
+	args := sort.args(key)
+	if store != "" {
+		args = append(args, "store", store)
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SortInterfaces(ctx context.Context, key string, sort *Sort) *SliceCmd {
+	cmd := NewSliceCmd(ctx, sort.args(key)...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Touch(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, len(keys)+1)
+	args[0] = "touch"
+	for i, key := range keys {
+		args[i+1] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) TTL(ctx context.Context, key string) *DurationCmd {
+	cmd := NewDurationCmd(ctx, time.Second, "ttl", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Type(ctx context.Context, key string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "type", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Append(ctx context.Context, key, value string) *IntCmd {
+	cmd := NewIntCmd(ctx, "append", key, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Decr(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "decr", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) DecrBy(ctx context.Context, key string, decrement int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "decrby", key, decrement)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `GET key` command. It returns redis.Nil error when key does not exist.
+func (c cmdable) Get(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "get", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) GetRange(ctx context.Context, key string, start, end int64) *StringCmd {
+	cmd := NewStringCmd(ctx, "getrange", key, start, end)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) GetSet(ctx context.Context, key string, value interface{}) *StringCmd {
+	cmd := NewStringCmd(ctx, "getset", key, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Incr(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "incr", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) IncrBy(ctx context.Context, key string, value int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "incrby", key, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) IncrByFloat(ctx context.Context, key string, value float64) *FloatCmd {
+	cmd := NewFloatCmd(ctx, "incrbyfloat", key, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) MGet(ctx context.Context, keys ...string) *SliceCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "mget"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// MSet is like Set but accepts multiple values:
+//   - MSet("key1", "value1", "key2", "value2")
+//   - MSet([]string{"key1", "value1", "key2", "value2"})
+//   - MSet(map[string]interface{}{"key1": "value1", "key2": "value2"})
+func (c cmdable) MSet(ctx context.Context, values ...interface{}) *StatusCmd {
+	args := make([]interface{}, 1, 1+len(values))
+	args[0] = "mset"
+	args = appendArgs(args, values)
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// MSetNX is like SetNX but accepts multiple values:
+//   - MSetNX("key1", "value1", "key2", "value2")
+//   - MSetNX([]string{"key1", "value1", "key2", "value2"})
+//   - MSetNX(map[string]interface{}{"key1": "value1", "key2": "value2"})
+func (c cmdable) MSetNX(ctx context.Context, values ...interface{}) *BoolCmd {
+	args := make([]interface{}, 1, 1+len(values))
+	args[0] = "msetnx"
+	args = appendArgs(args, values)
+	cmd := NewBoolCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SET key value [expiration]` command.
+// Use expiration for `SETEX`-like behavior.
+//
+// Zero expiration means the key has no expiration time.
+// KeepTTL(-1) expiration is a Redis KEEPTTL option to keep existing TTL.
+func (c cmdable) Set(ctx context.Context, key string, value interface{}, expiration time.Duration) *StatusCmd {
+	args := make([]interface{}, 3, 5)
+	args[0] = "set"
+	args[1] = key
+	args[2] = value
+	if expiration > 0 {
+		if usePrecise(expiration) {
+			args = append(args, "px", formatMs(ctx, expiration))
+		} else {
+			args = append(args, "ex", formatSec(ctx, expiration))
+		}
+	} else if expiration == KeepTTL {
+		args = append(args, "keepttl")
+	}
+
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SETEX key expiration value` command.
+func (c cmdable) SetEX(ctx context.Context, key string, value interface{}, expiration time.Duration) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "setex", key, formatSec(ctx, expiration), value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SET key value [expiration] NX` command.
+//
+// Zero expiration means the key has no expiration time.
+// KeepTTL(-1) expiration is a Redis KEEPTTL option to keep existing TTL.
+func (c cmdable) SetNX(ctx context.Context, key string, value interface{}, expiration time.Duration) *BoolCmd {
+	var cmd *BoolCmd
+	switch expiration {
+	case 0:
+		// Use old `SETNX` to support old Redis versions.
+		cmd = NewBoolCmd(ctx, "setnx", key, value)
+	case KeepTTL:
+		cmd = NewBoolCmd(ctx, "set", key, value, "keepttl", "nx")
+	default:
+		if usePrecise(expiration) {
+			cmd = NewBoolCmd(ctx, "set", key, value, "px", formatMs(ctx, expiration), "nx")
+		} else {
+			cmd = NewBoolCmd(ctx, "set", key, value, "ex", formatSec(ctx, expiration), "nx")
+		}
+	}
+
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SET key value [expiration] XX` command.
+//
+// Zero expiration means the key has no expiration time.
+// KeepTTL(-1) expiration is a Redis KEEPTTL option to keep existing TTL.
+func (c cmdable) SetXX(ctx context.Context, key string, value interface{}, expiration time.Duration) *BoolCmd {
+	var cmd *BoolCmd
+	switch expiration {
+	case 0:
+		cmd = NewBoolCmd(ctx, "set", key, value, "xx")
+	case KeepTTL:
+		cmd = NewBoolCmd(ctx, "set", key, value, "keepttl", "xx")
+	default:
+		if usePrecise(expiration) {
+			cmd = NewBoolCmd(ctx, "set", key, value, "px", formatMs(ctx, expiration), "xx")
+		} else {
+			cmd = NewBoolCmd(ctx, "set", key, value, "ex", formatSec(ctx, expiration), "xx")
+		}
+	}
+
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SetRange(ctx context.Context, key string, offset int64, value string) *IntCmd {
+	cmd := NewIntCmd(ctx, "setrange", key, offset, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) StrLen(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "strlen", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) GetBit(ctx context.Context, key string, offset int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "getbit", key, offset)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SetBit(ctx context.Context, key string, offset int64, value int) *IntCmd {
+	cmd := NewIntCmd(
+		ctx,
+		"setbit",
+		key,
+		offset,
+		value,
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type BitCount struct {
+	Start, End int64
+}
+
+func (c cmdable) BitCount(ctx context.Context, key string, bitCount *BitCount) *IntCmd {
+	args := []interface{}{"bitcount", key}
+	if bitCount != nil {
+		args = append(
+			args,
+			bitCount.Start,
+			bitCount.End,
+		)
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) bitOp(ctx context.Context, op, destKey string, keys ...string) *IntCmd {
+	args := make([]interface{}, 3+len(keys))
+	args[0] = "bitop"
+	args[1] = op
+	args[2] = destKey
+	for i, key := range keys {
+		args[3+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) BitOpAnd(ctx context.Context, destKey string, keys ...string) *IntCmd {
+	return c.bitOp(ctx, "and", destKey, keys...)
+}
+
+func (c cmdable) BitOpOr(ctx context.Context, destKey string, keys ...string) *IntCmd {
+	return c.bitOp(ctx, "or", destKey, keys...)
+}
+
+func (c cmdable) BitOpXor(ctx context.Context, destKey string, keys ...string) *IntCmd {
+	return c.bitOp(ctx, "xor", destKey, keys...)
+}
+
+func (c cmdable) BitOpNot(ctx context.Context, destKey string, key string) *IntCmd {
+	return c.bitOp(ctx, "not", destKey, key)
+}
+
+func (c cmdable) BitPos(ctx context.Context, key string, bit int64, pos ...int64) *IntCmd {
+	args := make([]interface{}, 3+len(pos))
+	args[0] = "bitpos"
+	args[1] = key
+	args[2] = bit
+	switch len(pos) {
+	case 0:
+	case 1:
+		args[3] = pos[0]
+	case 2:
+		args[3] = pos[0]
+		args[4] = pos[1]
+	default:
+		panic("too many arguments")
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) BitField(ctx context.Context, key string, args ...interface{}) *IntSliceCmd {
+	a := make([]interface{}, 0, 2+len(args))
+	a = append(a, "bitfield")
+	a = append(a, key)
+	a = append(a, args...)
+	cmd := NewIntSliceCmd(ctx, a...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) Scan(ctx context.Context, cursor uint64, match string, count int64) *ScanCmd {
+	args := []interface{}{"scan", cursor}
+	if match != "" {
+		args = append(args, "match", match)
+	}
+	if count > 0 {
+		args = append(args, "count", count)
+	}
+	cmd := NewScanCmd(ctx, c, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd {
+	args := []interface{}{"sscan", key, cursor}
+	if match != "" {
+		args = append(args, "match", match)
+	}
+	if count > 0 {
+		args = append(args, "count", count)
+	}
+	cmd := NewScanCmd(ctx, c, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd {
+	args := []interface{}{"hscan", key, cursor}
+	if match != "" {
+		args = append(args, "match", match)
+	}
+	if count > 0 {
+		args = append(args, "count", count)
+	}
+	cmd := NewScanCmd(ctx, c, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd {
+	args := []interface{}{"zscan", key, cursor}
+	if match != "" {
+		args = append(args, "match", match)
+	}
+	if count > 0 {
+		args = append(args, "count", count)
+	}
+	cmd := NewScanCmd(ctx, c, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) HDel(ctx context.Context, key string, fields ...string) *IntCmd {
+	args := make([]interface{}, 2+len(fields))
+	args[0] = "hdel"
+	args[1] = key
+	for i, field := range fields {
+		args[2+i] = field
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HExists(ctx context.Context, key, field string) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "hexists", key, field)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HGet(ctx context.Context, key, field string) *StringCmd {
+	cmd := NewStringCmd(ctx, "hget", key, field)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HGetAll(ctx context.Context, key string) *StringStringMapCmd {
+	cmd := NewStringStringMapCmd(ctx, "hgetall", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HIncrBy(ctx context.Context, key, field string, incr int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "hincrby", key, field, incr)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HIncrByFloat(ctx context.Context, key, field string, incr float64) *FloatCmd {
+	cmd := NewFloatCmd(ctx, "hincrbyfloat", key, field, incr)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HKeys(ctx context.Context, key string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "hkeys", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HLen(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "hlen", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// HMGet returns the values for the specified fields in the hash stored at key.
+// It returns an interface{} to distinguish between empty string and nil value.
+func (c cmdable) HMGet(ctx context.Context, key string, fields ...string) *SliceCmd {
+	args := make([]interface{}, 2+len(fields))
+	args[0] = "hmget"
+	args[1] = key
+	for i, field := range fields {
+		args[2+i] = field
+	}
+	cmd := NewSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// HSet accepts values in following formats:
+//   - HSet("myhash", "key1", "value1", "key2", "value2")
+//   - HSet("myhash", []string{"key1", "value1", "key2", "value2"})
+//   - HSet("myhash", map[string]interface{}{"key1": "value1", "key2": "value2"})
+//
+// Note that it requires Redis v4 for multiple field/value pairs support.
+func (c cmdable) HSet(ctx context.Context, key string, values ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "hset"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// HMSet is a deprecated version of HSet left for compatibility with Redis 3.
+func (c cmdable) HMSet(ctx context.Context, key string, values ...interface{}) *BoolCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "hmset"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewBoolCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HSetNX(ctx context.Context, key, field string, value interface{}) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "hsetnx", key, field, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) HVals(ctx context.Context, key string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "hvals", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) BLPop(ctx context.Context, timeout time.Duration, keys ...string) *StringSliceCmd {
+	args := make([]interface{}, 1+len(keys)+1)
+	args[0] = "blpop"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	args[len(args)-1] = formatSec(ctx, timeout)
+	cmd := NewStringSliceCmd(ctx, args...)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) BRPop(ctx context.Context, timeout time.Duration, keys ...string) *StringSliceCmd {
+	args := make([]interface{}, 1+len(keys)+1)
+	args[0] = "brpop"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	args[len(keys)+1] = formatSec(ctx, timeout)
+	cmd := NewStringSliceCmd(ctx, args...)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) BRPopLPush(ctx context.Context, source, destination string, timeout time.Duration) *StringCmd {
+	cmd := NewStringCmd(
+		ctx,
+		"brpoplpush",
+		source,
+		destination,
+		formatSec(ctx, timeout),
+	)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LIndex(ctx context.Context, key string, index int64) *StringCmd {
+	cmd := NewStringCmd(ctx, "lindex", key, index)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LInsert(ctx context.Context, key, op string, pivot, value interface{}) *IntCmd {
+	cmd := NewIntCmd(ctx, "linsert", key, op, pivot, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LInsertBefore(ctx context.Context, key string, pivot, value interface{}) *IntCmd {
+	cmd := NewIntCmd(ctx, "linsert", key, "before", pivot, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LInsertAfter(ctx context.Context, key string, pivot, value interface{}) *IntCmd {
+	cmd := NewIntCmd(ctx, "linsert", key, "after", pivot, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LLen(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "llen", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LPop(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "lpop", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type LPosArgs struct {
+	Rank, MaxLen int64
+}
+
+func (c cmdable) LPos(ctx context.Context, key string, value string, a LPosArgs) *IntCmd {
+	args := []interface{}{"lpos", key, value}
+	if a.Rank != 0 {
+		args = append(args, "rank", a.Rank)
+	}
+	if a.MaxLen != 0 {
+		args = append(args, "maxlen", a.MaxLen)
+	}
+
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LPosCount(ctx context.Context, key string, value string, count int64, a LPosArgs) *IntSliceCmd {
+	args := []interface{}{"lpos", key, value, "count", count}
+	if a.Rank != 0 {
+		args = append(args, "rank", a.Rank)
+	}
+	if a.MaxLen != 0 {
+		args = append(args, "maxlen", a.MaxLen)
+	}
+	cmd := NewIntSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LPush(ctx context.Context, key string, values ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "lpush"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LPushX(ctx context.Context, key string, values ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "lpushx"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd {
+	cmd := NewStringSliceCmd(
+		ctx,
+		"lrange",
+		key,
+		start,
+		stop,
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LRem(ctx context.Context, key string, count int64, value interface{}) *IntCmd {
+	cmd := NewIntCmd(ctx, "lrem", key, count, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LSet(ctx context.Context, key string, index int64, value interface{}) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "lset", key, index, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LTrim(ctx context.Context, key string, start, stop int64) *StatusCmd {
+	cmd := NewStatusCmd(
+		ctx,
+		"ltrim",
+		key,
+		start,
+		stop,
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RPop(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "rpop", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RPopLPush(ctx context.Context, source, destination string) *StringCmd {
+	cmd := NewStringCmd(ctx, "rpoplpush", source, destination)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RPush(ctx context.Context, key string, values ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "rpush"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) RPushX(ctx context.Context, key string, values ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(values))
+	args[0] = "rpushx"
+	args[1] = key
+	args = appendArgs(args, values)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) SAdd(ctx context.Context, key string, members ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(members))
+	args[0] = "sadd"
+	args[1] = key
+	args = appendArgs(args, members)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SCard(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "scard", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SDiff(ctx context.Context, keys ...string) *StringSliceCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "sdiff"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SDiffStore(ctx context.Context, destination string, keys ...string) *IntCmd {
+	args := make([]interface{}, 2+len(keys))
+	args[0] = "sdiffstore"
+	args[1] = destination
+	for i, key := range keys {
+		args[2+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SInter(ctx context.Context, keys ...string) *StringSliceCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "sinter"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SInterStore(ctx context.Context, destination string, keys ...string) *IntCmd {
+	args := make([]interface{}, 2+len(keys))
+	args[0] = "sinterstore"
+	args[1] = destination
+	for i, key := range keys {
+		args[2+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SIsMember(ctx context.Context, key string, member interface{}) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "sismember", key, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SMEMBERS key` command output as a slice.
+func (c cmdable) SMembers(ctx context.Context, key string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "smembers", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SMEMBERS key` command output as a map.
+func (c cmdable) SMembersMap(ctx context.Context, key string) *StringStructMapCmd {
+	cmd := NewStringStructMapCmd(ctx, "smembers", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SMove(ctx context.Context, source, destination string, member interface{}) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "smove", source, destination, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SPOP key` command.
+func (c cmdable) SPop(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "spop", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SPOP key count` command.
+func (c cmdable) SPopN(ctx context.Context, key string, count int64) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "spop", key, count)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SRANDMEMBER key` command.
+func (c cmdable) SRandMember(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "srandmember", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `SRANDMEMBER key count` command.
+func (c cmdable) SRandMemberN(ctx context.Context, key string, count int64) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "srandmember", key, count)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SRem(ctx context.Context, key string, members ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(members))
+	args[0] = "srem"
+	args[1] = key
+	args = appendArgs(args, members)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SUnion(ctx context.Context, keys ...string) *StringSliceCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "sunion"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SUnionStore(ctx context.Context, destination string, keys ...string) *IntCmd {
+	args := make([]interface{}, 2+len(keys))
+	args[0] = "sunionstore"
+	args[1] = destination
+	for i, key := range keys {
+		args[2+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+// XAddArgs accepts values in the following formats:
+//   - XAddArgs.Values = []interface{}{"key1", "value1", "key2", "value2"}
+//   - XAddArgs.Values = []string("key1", "value1", "key2", "value2")
+//   - XAddArgs.Values = map[string]interface{}{"key1": "value1", "key2": "value2"}
+//
+// Note that map will not preserve the order of key-value pairs.
+type XAddArgs struct {
+	Stream       string
+	MaxLen       int64 // MAXLEN N
+	MaxLenApprox int64 // MAXLEN ~ N
+	ID           string
+	Values       interface{}
+}
+
+func (c cmdable) XAdd(ctx context.Context, a *XAddArgs) *StringCmd {
+	args := make([]interface{}, 0, 8)
+	args = append(args, "xadd")
+	args = append(args, a.Stream)
+	if a.MaxLen > 0 {
+		args = append(args, "maxlen", a.MaxLen)
+	} else if a.MaxLenApprox > 0 {
+		args = append(args, "maxlen", "~", a.MaxLenApprox)
+	}
+	if a.ID != "" {
+		args = append(args, a.ID)
+	} else {
+		args = append(args, "*")
+	}
+	args = appendArg(args, a.Values)
+
+	cmd := NewStringCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XDel(ctx context.Context, stream string, ids ...string) *IntCmd {
+	args := []interface{}{"xdel", stream}
+	for _, id := range ids {
+		args = append(args, id)
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XLen(ctx context.Context, stream string) *IntCmd {
+	cmd := NewIntCmd(ctx, "xlen", stream)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XRange(ctx context.Context, stream, start, stop string) *XMessageSliceCmd {
+	cmd := NewXMessageSliceCmd(ctx, "xrange", stream, start, stop)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XRangeN(ctx context.Context, stream, start, stop string, count int64) *XMessageSliceCmd {
+	cmd := NewXMessageSliceCmd(ctx, "xrange", stream, start, stop, "count", count)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XRevRange(ctx context.Context, stream, start, stop string) *XMessageSliceCmd {
+	cmd := NewXMessageSliceCmd(ctx, "xrevrange", stream, start, stop)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XRevRangeN(ctx context.Context, stream, start, stop string, count int64) *XMessageSliceCmd {
+	cmd := NewXMessageSliceCmd(ctx, "xrevrange", stream, start, stop, "count", count)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type XReadArgs struct {
+	Streams []string // list of streams and ids, e.g. stream1 stream2 id1 id2
+	Count   int64
+	Block   time.Duration
+}
+
+func (c cmdable) XRead(ctx context.Context, a *XReadArgs) *XStreamSliceCmd {
+	args := make([]interface{}, 0, 5+len(a.Streams))
+	args = append(args, "xread")
+
+	keyPos := int8(1)
+	if a.Count > 0 {
+		args = append(args, "count")
+		args = append(args, a.Count)
+		keyPos += 2
+	}
+	if a.Block >= 0 {
+		args = append(args, "block")
+		args = append(args, int64(a.Block/time.Millisecond))
+		keyPos += 2
+	}
+	args = append(args, "streams")
+	keyPos++
+	for _, s := range a.Streams {
+		args = append(args, s)
+	}
+
+	cmd := NewXStreamSliceCmd(ctx, args...)
+	if a.Block >= 0 {
+		cmd.setReadTimeout(a.Block)
+	}
+	cmd.setFirstKeyPos(keyPos)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XReadStreams(ctx context.Context, streams ...string) *XStreamSliceCmd {
+	return c.XRead(ctx, &XReadArgs{
+		Streams: streams,
+		Block:   -1,
+	})
+}
+
+func (c cmdable) XGroupCreate(ctx context.Context, stream, group, start string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "xgroup", "create", stream, group, start)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XGroupCreateMkStream(ctx context.Context, stream, group, start string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "xgroup", "create", stream, group, start, "mkstream")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XGroupSetID(ctx context.Context, stream, group, start string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "xgroup", "setid", stream, group, start)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XGroupDestroy(ctx context.Context, stream, group string) *IntCmd {
+	cmd := NewIntCmd(ctx, "xgroup", "destroy", stream, group)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XGroupDelConsumer(ctx context.Context, stream, group, consumer string) *IntCmd {
+	cmd := NewIntCmd(ctx, "xgroup", "delconsumer", stream, group, consumer)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type XReadGroupArgs struct {
+	Group    string
+	Consumer string
+	Streams  []string // list of streams and ids, e.g. stream1 stream2 id1 id2
+	Count    int64
+	Block    time.Duration
+	NoAck    bool
+}
+
+func (c cmdable) XReadGroup(ctx context.Context, a *XReadGroupArgs) *XStreamSliceCmd {
+	args := make([]interface{}, 0, 8+len(a.Streams))
+	args = append(args, "xreadgroup", "group", a.Group, a.Consumer)
+
+	keyPos := int8(1)
+	if a.Count > 0 {
+		args = append(args, "count", a.Count)
+		keyPos += 2
+	}
+	if a.Block >= 0 {
+		args = append(args, "block", int64(a.Block/time.Millisecond))
+		keyPos += 2
+	}
+	if a.NoAck {
+		args = append(args, "noack")
+		keyPos++
+	}
+	args = append(args, "streams")
+	keyPos++
+	for _, s := range a.Streams {
+		args = append(args, s)
+	}
+
+	cmd := NewXStreamSliceCmd(ctx, args...)
+	if a.Block >= 0 {
+		cmd.setReadTimeout(a.Block)
+	}
+	cmd.setFirstKeyPos(keyPos)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XAck(ctx context.Context, stream, group string, ids ...string) *IntCmd {
+	args := []interface{}{"xack", stream, group}
+	for _, id := range ids {
+		args = append(args, id)
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XPending(ctx context.Context, stream, group string) *XPendingCmd {
+	cmd := NewXPendingCmd(ctx, "xpending", stream, group)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type XPendingExtArgs struct {
+	Stream   string
+	Group    string
+	Start    string
+	End      string
+	Count    int64
+	Consumer string
+}
+
+func (c cmdable) XPendingExt(ctx context.Context, a *XPendingExtArgs) *XPendingExtCmd {
+	args := make([]interface{}, 0, 7)
+	args = append(args, "xpending", a.Stream, a.Group, a.Start, a.End, a.Count)
+	if a.Consumer != "" {
+		args = append(args, a.Consumer)
+	}
+	cmd := NewXPendingExtCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type XClaimArgs struct {
+	Stream   string
+	Group    string
+	Consumer string
+	MinIdle  time.Duration
+	Messages []string
+}
+
+func (c cmdable) XClaim(ctx context.Context, a *XClaimArgs) *XMessageSliceCmd {
+	args := xClaimArgs(a)
+	cmd := NewXMessageSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XClaimJustID(ctx context.Context, a *XClaimArgs) *StringSliceCmd {
+	args := xClaimArgs(a)
+	args = append(args, "justid")
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func xClaimArgs(a *XClaimArgs) []interface{} {
+	args := make([]interface{}, 0, 4+len(a.Messages))
+	args = append(args,
+		"xclaim",
+		a.Stream,
+		a.Group, a.Consumer,
+		int64(a.MinIdle/time.Millisecond))
+	for _, id := range a.Messages {
+		args = append(args, id)
+	}
+	return args
+}
+
+func (c cmdable) XTrim(ctx context.Context, key string, maxLen int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "xtrim", key, "maxlen", maxLen)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XTrimApprox(ctx context.Context, key string, maxLen int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "xtrim", key, "maxlen", "~", maxLen)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XInfoGroups(ctx context.Context, key string) *XInfoGroupsCmd {
+	cmd := NewXInfoGroupsCmd(ctx, key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) XInfoStream(ctx context.Context, key string) *XInfoStreamCmd {
+	cmd := NewXInfoStreamCmd(ctx, key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+// Z represents sorted set member.
+type Z struct {
+	Score  float64
+	Member interface{}
+}
+
+// ZWithKey represents sorted set member including the name of the key where it was popped.
+type ZWithKey struct {
+	Z
+	Key string
+}
+
+// ZStore is used as an arg to ZInterStore and ZUnionStore.
+type ZStore struct {
+	Keys    []string
+	Weights []float64
+	// Can be SUM, MIN or MAX.
+	Aggregate string
+}
+
+// Redis `BZPOPMAX key [key ...] timeout` command.
+func (c cmdable) BZPopMax(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd {
+	args := make([]interface{}, 1+len(keys)+1)
+	args[0] = "bzpopmax"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	args[len(args)-1] = formatSec(ctx, timeout)
+	cmd := NewZWithKeyCmd(ctx, args...)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `BZPOPMIN key [key ...] timeout` command.
+func (c cmdable) BZPopMin(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd {
+	args := make([]interface{}, 1+len(keys)+1)
+	args[0] = "bzpopmin"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	args[len(args)-1] = formatSec(ctx, timeout)
+	cmd := NewZWithKeyCmd(ctx, args...)
+	cmd.setReadTimeout(timeout)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) zAdd(ctx context.Context, a []interface{}, n int, members ...*Z) *IntCmd {
+	for i, m := range members {
+		a[n+2*i] = m.Score
+		a[n+2*i+1] = m.Member
+	}
+	cmd := NewIntCmd(ctx, a...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `ZADD key score member [score member ...]` command.
+func (c cmdable) ZAdd(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 2
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1] = "zadd", key
+	return c.zAdd(ctx, a, n, members...)
+}
+
+// Redis `ZADD key NX score member [score member ...]` command.
+func (c cmdable) ZAddNX(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 3
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1], a[2] = "zadd", key, "nx"
+	return c.zAdd(ctx, a, n, members...)
+}
+
+// Redis `ZADD key XX score member [score member ...]` command.
+func (c cmdable) ZAddXX(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 3
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1], a[2] = "zadd", key, "xx"
+	return c.zAdd(ctx, a, n, members...)
+}
+
+// Redis `ZADD key CH score member [score member ...]` command.
+func (c cmdable) ZAddCh(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 3
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1], a[2] = "zadd", key, "ch"
+	return c.zAdd(ctx, a, n, members...)
+}
+
+// Redis `ZADD key NX CH score member [score member ...]` command.
+func (c cmdable) ZAddNXCh(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 4
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1], a[2], a[3] = "zadd", key, "nx", "ch"
+	return c.zAdd(ctx, a, n, members...)
+}
+
+// Redis `ZADD key XX CH score member [score member ...]` command.
+func (c cmdable) ZAddXXCh(ctx context.Context, key string, members ...*Z) *IntCmd {
+	const n = 4
+	a := make([]interface{}, n+2*len(members))
+	a[0], a[1], a[2], a[3] = "zadd", key, "xx", "ch"
+	return c.zAdd(ctx, a, n, members...)
+}
+
+func (c cmdable) zIncr(ctx context.Context, a []interface{}, n int, members ...*Z) *FloatCmd {
+	for i, m := range members {
+		a[n+2*i] = m.Score
+		a[n+2*i+1] = m.Member
+	}
+	cmd := NewFloatCmd(ctx, a...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// Redis `ZADD key INCR score member` command.
+func (c cmdable) ZIncr(ctx context.Context, key string, member *Z) *FloatCmd {
+	const n = 3
+	a := make([]interface{}, n+2)
+	a[0], a[1], a[2] = "zadd", key, "incr"
+	return c.zIncr(ctx, a, n, member)
+}
+
+// Redis `ZADD key NX INCR score member` command.
+func (c cmdable) ZIncrNX(ctx context.Context, key string, member *Z) *FloatCmd {
+	const n = 4
+	a := make([]interface{}, n+2)
+	a[0], a[1], a[2], a[3] = "zadd", key, "incr", "nx"
+	return c.zIncr(ctx, a, n, member)
+}
+
+// Redis `ZADD key XX INCR score member` command.
+func (c cmdable) ZIncrXX(ctx context.Context, key string, member *Z) *FloatCmd {
+	const n = 4
+	a := make([]interface{}, n+2)
+	a[0], a[1], a[2], a[3] = "zadd", key, "incr", "xx"
+	return c.zIncr(ctx, a, n, member)
+}
+
+func (c cmdable) ZCard(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zcard", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZCount(ctx context.Context, key, min, max string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zcount", key, min, max)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZLexCount(ctx context.Context, key, min, max string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zlexcount", key, min, max)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZIncrBy(ctx context.Context, key string, increment float64, member string) *FloatCmd {
+	cmd := NewFloatCmd(ctx, "zincrby", key, increment, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZInterStore(ctx context.Context, destination string, store *ZStore) *IntCmd {
+	args := make([]interface{}, 3+len(store.Keys))
+	args[0] = "zinterstore"
+	args[1] = destination
+	args[2] = len(store.Keys)
+	for i, key := range store.Keys {
+		args[3+i] = key
+	}
+	if len(store.Weights) > 0 {
+		args = append(args, "weights")
+		for _, weight := range store.Weights {
+			args = append(args, weight)
+		}
+	}
+	if store.Aggregate != "" {
+		args = append(args, "aggregate", store.Aggregate)
+	}
+	cmd := NewIntCmd(ctx, args...)
+	cmd.setFirstKeyPos(3)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZPopMax(ctx context.Context, key string, count ...int64) *ZSliceCmd {
+	args := []interface{}{
+		"zpopmax",
+		key,
+	}
+
+	switch len(count) {
+	case 0:
+		break
+	case 1:
+		args = append(args, count[0])
+	default:
+		panic("too many arguments")
+	}
+
+	cmd := NewZSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZPopMin(ctx context.Context, key string, count ...int64) *ZSliceCmd {
+	args := []interface{}{
+		"zpopmin",
+		key,
+	}
+
+	switch len(count) {
+	case 0:
+		break
+	case 1:
+		args = append(args, count[0])
+	default:
+		panic("too many arguments")
+	}
+
+	cmd := NewZSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) zRange(ctx context.Context, key string, start, stop int64, withScores bool) *StringSliceCmd {
+	args := []interface{}{
+		"zrange",
+		key,
+		start,
+		stop,
+	}
+	if withScores {
+		args = append(args, "withscores")
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd {
+	return c.zRange(ctx, key, start, stop, false)
+}
+
+func (c cmdable) ZRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd {
+	cmd := NewZSliceCmd(ctx, "zrange", key, start, stop, "withscores")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+type ZRangeBy struct {
+	Min, Max      string
+	Offset, Count int64
+}
+
+func (c cmdable) zRangeBy(ctx context.Context, zcmd, key string, opt *ZRangeBy, withScores bool) *StringSliceCmd {
+	args := []interface{}{zcmd, key, opt.Min, opt.Max}
+	if withScores {
+		args = append(args, "withscores")
+	}
+	if opt.Offset != 0 || opt.Count != 0 {
+		args = append(
+			args,
+			"limit",
+			opt.Offset,
+			opt.Count,
+		)
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
+	return c.zRangeBy(ctx, "zrangebyscore", key, opt, false)
+}
+
+func (c cmdable) ZRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
+	return c.zRangeBy(ctx, "zrangebylex", key, opt, false)
+}
+
+func (c cmdable) ZRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd {
+	args := []interface{}{"zrangebyscore", key, opt.Min, opt.Max, "withscores"}
+	if opt.Offset != 0 || opt.Count != 0 {
+		args = append(
+			args,
+			"limit",
+			opt.Offset,
+			opt.Count,
+		)
+	}
+	cmd := NewZSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRank(ctx context.Context, key, member string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zrank", key, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRem(ctx context.Context, key string, members ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(members))
+	args[0] = "zrem"
+	args[1] = key
+	args = appendArgs(args, members)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRemRangeByRank(ctx context.Context, key string, start, stop int64) *IntCmd {
+	cmd := NewIntCmd(
+		ctx,
+		"zremrangebyrank",
+		key,
+		start,
+		stop,
+	)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRemRangeByScore(ctx context.Context, key, min, max string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zremrangebyscore", key, min, max)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRemRangeByLex(ctx context.Context, key, min, max string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zremrangebylex", key, min, max)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRevRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "zrevrange", key, start, stop)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRevRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd {
+	cmd := NewZSliceCmd(ctx, "zrevrange", key, start, stop, "withscores")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) zRevRangeBy(ctx context.Context, zcmd, key string, opt *ZRangeBy) *StringSliceCmd {
+	args := []interface{}{zcmd, key, opt.Max, opt.Min}
+	if opt.Offset != 0 || opt.Count != 0 {
+		args = append(
+			args,
+			"limit",
+			opt.Offset,
+			opt.Count,
+		)
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRevRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
+	return c.zRevRangeBy(ctx, "zrevrangebyscore", key, opt)
+}
+
+func (c cmdable) ZRevRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
+	return c.zRevRangeBy(ctx, "zrevrangebylex", key, opt)
+}
+
+func (c cmdable) ZRevRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd {
+	args := []interface{}{"zrevrangebyscore", key, opt.Max, opt.Min, "withscores"}
+	if opt.Offset != 0 || opt.Count != 0 {
+		args = append(
+			args,
+			"limit",
+			opt.Offset,
+			opt.Count,
+		)
+	}
+	cmd := NewZSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZRevRank(ctx context.Context, key, member string) *IntCmd {
+	cmd := NewIntCmd(ctx, "zrevrank", key, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZScore(ctx context.Context, key, member string) *FloatCmd {
+	cmd := NewFloatCmd(ctx, "zscore", key, member)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ZUnionStore(ctx context.Context, dest string, store *ZStore) *IntCmd {
+	args := make([]interface{}, 3+len(store.Keys))
+	args[0] = "zunionstore"
+	args[1] = dest
+	args[2] = len(store.Keys)
+	for i, key := range store.Keys {
+		args[3+i] = key
+	}
+	if len(store.Weights) > 0 {
+		args = append(args, "weights")
+		for _, weight := range store.Weights {
+			args = append(args, weight)
+		}
+	}
+	if store.Aggregate != "" {
+		args = append(args, "aggregate", store.Aggregate)
+	}
+
+	cmd := NewIntCmd(ctx, args...)
+	cmd.setFirstKeyPos(3)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) PFAdd(ctx context.Context, key string, els ...interface{}) *IntCmd {
+	args := make([]interface{}, 2, 2+len(els))
+	args[0] = "pfadd"
+	args[1] = key
+	args = appendArgs(args, els)
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PFCount(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "pfcount"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PFMerge(ctx context.Context, dest string, keys ...string) *StatusCmd {
+	args := make([]interface{}, 2+len(keys))
+	args[0] = "pfmerge"
+	args[1] = dest
+	for i, key := range keys {
+		args[2+i] = key
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) BgRewriteAOF(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "bgrewriteaof")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) BgSave(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "bgsave")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientKill(ctx context.Context, ipPort string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "client", "kill", ipPort)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// ClientKillByFilter is new style syntax, while the ClientKill is old
+//
+//   CLIENT KILL <option> [value] ... <option> [value]
+func (c cmdable) ClientKillByFilter(ctx context.Context, keys ...string) *IntCmd {
+	args := make([]interface{}, 2+len(keys))
+	args[0] = "client"
+	args[1] = "kill"
+	for i, key := range keys {
+		args[2+i] = key
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientList(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "client", "list")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientPause(ctx context.Context, dur time.Duration) *BoolCmd {
+	cmd := NewBoolCmd(ctx, "client", "pause", formatMs(ctx, dur))
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientID(ctx context.Context) *IntCmd {
+	cmd := NewIntCmd(ctx, "client", "id")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientUnblock(ctx context.Context, id int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "client", "unblock", id)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClientUnblockWithError(ctx context.Context, id int64) *IntCmd {
+	cmd := NewIntCmd(ctx, "client", "unblock", id, "error")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ConfigGet(ctx context.Context, parameter string) *SliceCmd {
+	cmd := NewSliceCmd(ctx, "config", "get", parameter)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ConfigResetStat(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "config", "resetstat")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ConfigSet(ctx context.Context, parameter, value string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "config", "set", parameter, value)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ConfigRewrite(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "config", "rewrite")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) DBSize(ctx context.Context) *IntCmd {
+	cmd := NewIntCmd(ctx, "dbsize")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) FlushAll(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "flushall")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) FlushAllAsync(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "flushall", "async")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) FlushDB(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "flushdb")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) FlushDBAsync(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "flushdb", "async")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Info(ctx context.Context, section ...string) *StringCmd {
+	args := []interface{}{"info"}
+	if len(section) > 0 {
+		args = append(args, section[0])
+	}
+	cmd := NewStringCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) LastSave(ctx context.Context) *IntCmd {
+	cmd := NewIntCmd(ctx, "lastsave")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Save(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "save")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) shutdown(ctx context.Context, modifier string) *StatusCmd {
+	var args []interface{}
+	if modifier == "" {
+		args = []interface{}{"shutdown"}
+	} else {
+		args = []interface{}{"shutdown", modifier}
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	if err := cmd.Err(); err != nil {
+		if err == io.EOF {
+			// Server quit as expected.
+			cmd.err = nil
+		}
+	} else {
+		// Server did not quit. String reply contains the reason.
+		cmd.err = errors.New(cmd.val)
+		cmd.val = ""
+	}
+	return cmd
+}
+
+func (c cmdable) Shutdown(ctx context.Context) *StatusCmd {
+	return c.shutdown(ctx, "")
+}
+
+func (c cmdable) ShutdownSave(ctx context.Context) *StatusCmd {
+	return c.shutdown(ctx, "save")
+}
+
+func (c cmdable) ShutdownNoSave(ctx context.Context) *StatusCmd {
+	return c.shutdown(ctx, "nosave")
+}
+
+func (c cmdable) SlaveOf(ctx context.Context, host, port string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "slaveof", host, port)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) SlowLogGet(ctx context.Context, num int64) *SlowLogCmd {
+	cmd := NewSlowLogCmd(context.Background(), "slowlog", "get", num)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) Sync(ctx context.Context) {
+	panic("not implemented")
+}
+
+func (c cmdable) Time(ctx context.Context) *TimeCmd {
+	cmd := NewTimeCmd(ctx, "time")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) DebugObject(ctx context.Context, key string) *StringCmd {
+	cmd := NewStringCmd(ctx, "debug", "object", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ReadOnly(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "readonly")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ReadWrite(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "readwrite")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) MemoryUsage(ctx context.Context, key string, samples ...int) *IntCmd {
+	args := []interface{}{"memory", "usage", key}
+	if len(samples) > 0 {
+		if len(samples) != 1 {
+			panic("MemoryUsage expects single sample count")
+		}
+		args = append(args, "SAMPLES", samples[0])
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) Eval(ctx context.Context, script string, keys []string, args ...interface{}) *Cmd {
+	cmdArgs := make([]interface{}, 3+len(keys), 3+len(keys)+len(args))
+	cmdArgs[0] = "eval"
+	cmdArgs[1] = script
+	cmdArgs[2] = len(keys)
+	for i, key := range keys {
+		cmdArgs[3+i] = key
+	}
+	cmdArgs = appendArgs(cmdArgs, args)
+	cmd := NewCmd(ctx, cmdArgs...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) EvalSha(ctx context.Context, sha1 string, keys []string, args ...interface{}) *Cmd {
+	cmdArgs := make([]interface{}, 3+len(keys), 3+len(keys)+len(args))
+	cmdArgs[0] = "evalsha"
+	cmdArgs[1] = sha1
+	cmdArgs[2] = len(keys)
+	for i, key := range keys {
+		cmdArgs[3+i] = key
+	}
+	cmdArgs = appendArgs(cmdArgs, args)
+	cmd := NewCmd(ctx, cmdArgs...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ScriptExists(ctx context.Context, hashes ...string) *BoolSliceCmd {
+	args := make([]interface{}, 2+len(hashes))
+	args[0] = "script"
+	args[1] = "exists"
+	for i, hash := range hashes {
+		args[2+i] = hash
+	}
+	cmd := NewBoolSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ScriptFlush(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "script", "flush")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ScriptKill(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "script", "kill")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ScriptLoad(ctx context.Context, script string) *StringCmd {
+	cmd := NewStringCmd(ctx, "script", "load", script)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+// Publish posts the message to the channel.
+func (c cmdable) Publish(ctx context.Context, channel string, message interface{}) *IntCmd {
+	cmd := NewIntCmd(ctx, "publish", channel, message)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PubSubChannels(ctx context.Context, pattern string) *StringSliceCmd {
+	args := []interface{}{"pubsub", "channels"}
+	if pattern != "*" {
+		args = append(args, pattern)
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PubSubNumSub(ctx context.Context, channels ...string) *StringIntMapCmd {
+	args := make([]interface{}, 2+len(channels))
+	args[0] = "pubsub"
+	args[1] = "numsub"
+	for i, channel := range channels {
+		args[2+i] = channel
+	}
+	cmd := NewStringIntMapCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) PubSubNumPat(ctx context.Context) *IntCmd {
+	cmd := NewIntCmd(ctx, "pubsub", "numpat")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) ClusterSlots(ctx context.Context) *ClusterSlotsCmd {
+	cmd := NewClusterSlotsCmd(ctx, "cluster", "slots")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterNodes(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "cluster", "nodes")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterMeet(ctx context.Context, host, port string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "meet", host, port)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterForget(ctx context.Context, nodeID string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "forget", nodeID)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterReplicate(ctx context.Context, nodeID string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "replicate", nodeID)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterResetSoft(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "reset", "soft")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterResetHard(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "reset", "hard")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterInfo(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "cluster", "info")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterKeySlot(ctx context.Context, key string) *IntCmd {
+	cmd := NewIntCmd(ctx, "cluster", "keyslot", key)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterGetKeysInSlot(ctx context.Context, slot int, count int) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "cluster", "getkeysinslot", slot, count)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterCountFailureReports(ctx context.Context, nodeID string) *IntCmd {
+	cmd := NewIntCmd(ctx, "cluster", "count-failure-reports", nodeID)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterCountKeysInSlot(ctx context.Context, slot int) *IntCmd {
+	cmd := NewIntCmd(ctx, "cluster", "countkeysinslot", slot)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterDelSlots(ctx context.Context, slots ...int) *StatusCmd {
+	args := make([]interface{}, 2+len(slots))
+	args[0] = "cluster"
+	args[1] = "delslots"
+	for i, slot := range slots {
+		args[2+i] = slot
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterDelSlotsRange(ctx context.Context, min, max int) *StatusCmd {
+	size := max - min + 1
+	slots := make([]int, size)
+	for i := 0; i < size; i++ {
+		slots[i] = min + i
+	}
+	return c.ClusterDelSlots(ctx, slots...)
+}
+
+func (c cmdable) ClusterSaveConfig(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "saveconfig")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterSlaves(ctx context.Context, nodeID string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "cluster", "slaves", nodeID)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterFailover(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "cluster", "failover")
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterAddSlots(ctx context.Context, slots ...int) *StatusCmd {
+	args := make([]interface{}, 2+len(slots))
+	args[0] = "cluster"
+	args[1] = "addslots"
+	for i, num := range slots {
+		args[2+i] = num
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) ClusterAddSlotsRange(ctx context.Context, min, max int) *StatusCmd {
+	size := max - min + 1
+	slots := make([]int, size)
+	for i := 0; i < size; i++ {
+		slots[i] = min + i
+	}
+	return c.ClusterAddSlots(ctx, slots...)
+}
+
+//------------------------------------------------------------------------------
+
+func (c cmdable) GeoAdd(ctx context.Context, key string, geoLocation ...*GeoLocation) *IntCmd {
+	args := make([]interface{}, 2+3*len(geoLocation))
+	args[0] = "geoadd"
+	args[1] = key
+	for i, eachLoc := range geoLocation {
+		args[2+3*i] = eachLoc.Longitude
+		args[2+3*i+1] = eachLoc.Latitude
+		args[2+3*i+2] = eachLoc.Name
+	}
+	cmd := NewIntCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// GeoRadius is a read-only GEORADIUS_RO command.
+func (c cmdable) GeoRadius(
+	ctx context.Context, key string, longitude, latitude float64, query *GeoRadiusQuery,
+) *GeoLocationCmd {
+	cmd := NewGeoLocationCmd(ctx, query, "georadius_ro", key, longitude, latitude)
+	if query.Store != "" || query.StoreDist != "" {
+		cmd.SetErr(errors.New("GeoRadius does not support Store or StoreDist"))
+		return cmd
+	}
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// GeoRadiusStore is a writing GEORADIUS command.
+func (c cmdable) GeoRadiusStore(
+	ctx context.Context, key string, longitude, latitude float64, query *GeoRadiusQuery,
+) *IntCmd {
+	args := geoLocationArgs(query, "georadius", key, longitude, latitude)
+	cmd := NewIntCmd(ctx, args...)
+	if query.Store == "" && query.StoreDist == "" {
+		cmd.SetErr(errors.New("GeoRadiusStore requires Store or StoreDist"))
+		return cmd
+	}
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// GeoRadius is a read-only GEORADIUSBYMEMBER_RO command.
+func (c cmdable) GeoRadiusByMember(
+	ctx context.Context, key, member string, query *GeoRadiusQuery,
+) *GeoLocationCmd {
+	cmd := NewGeoLocationCmd(ctx, query, "georadiusbymember_ro", key, member)
+	if query.Store != "" || query.StoreDist != "" {
+		cmd.SetErr(errors.New("GeoRadiusByMember does not support Store or StoreDist"))
+		return cmd
+	}
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+// GeoRadiusByMemberStore is a writing GEORADIUSBYMEMBER command.
+func (c cmdable) GeoRadiusByMemberStore(
+	ctx context.Context, key, member string, query *GeoRadiusQuery,
+) *IntCmd {
+	args := geoLocationArgs(query, "georadiusbymember", key, member)
+	cmd := NewIntCmd(ctx, args...)
+	if query.Store == "" && query.StoreDist == "" {
+		cmd.SetErr(errors.New("GeoRadiusByMemberStore requires Store or StoreDist"))
+		return cmd
+	}
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) GeoDist(
+	ctx context.Context, key string, member1, member2, unit string,
+) *FloatCmd {
+	if unit == "" {
+		unit = "km"
+	}
+	cmd := NewFloatCmd(ctx, "geodist", key, member1, member2, unit)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) GeoHash(ctx context.Context, key string, members ...string) *StringSliceCmd {
+	args := make([]interface{}, 2+len(members))
+	args[0] = "geohash"
+	args[1] = key
+	for i, member := range members {
+		args[2+i] = member
+	}
+	cmd := NewStringSliceCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
+
+func (c cmdable) GeoPos(ctx context.Context, key string, members ...string) *GeoPosCmd {
+	args := make([]interface{}, 2+len(members))
+	args[0] = "geopos"
+	args[1] = key
+	for i, member := range members {
+		args[2+i] = member
+	}
+	cmd := NewGeoPosCmd(ctx, args...)
+	_ = c(ctx, cmd)
+	return cmd
+}
diff --git a/vendor/github.com/go-redis/redis/v8/doc.go b/vendor/github.com/go-redis/redis/v8/doc.go
new file mode 100644
index 0000000..5526253
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/doc.go
@@ -0,0 +1,4 @@
+/*
+Package redis implements a Redis client.
+*/
+package redis
diff --git a/vendor/github.com/go-redis/redis/v8/error.go b/vendor/github.com/go-redis/redis/v8/error.go
new file mode 100644
index 0000000..9fe1376
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/error.go
@@ -0,0 +1,122 @@
+package redis
+
+import (
+	"context"
+	"io"
+	"net"
+	"strings"
+
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/proto"
+)
+
+var ErrClosed = pool.ErrClosed
+
+type Error interface {
+	error
+
+	// RedisError is a no-op function but
+	// serves to distinguish types that are Redis
+	// errors from ordinary errors: a type is a
+	// Redis error if it has a RedisError method.
+	RedisError()
+}
+
+var _ Error = proto.RedisError("")
+
+func shouldRetry(err error, retryTimeout bool) bool {
+	switch err {
+	case io.EOF, io.ErrUnexpectedEOF:
+		return true
+	case nil, context.Canceled, context.DeadlineExceeded:
+		return false
+	}
+
+	if v, ok := err.(timeoutError); ok {
+		if v.Timeout() {
+			return retryTimeout
+		}
+		return true
+	}
+
+	s := err.Error()
+	if s == "ERR max number of clients reached" {
+		return true
+	}
+	if strings.HasPrefix(s, "LOADING ") {
+		return true
+	}
+	if strings.HasPrefix(s, "READONLY ") {
+		return true
+	}
+	if strings.HasPrefix(s, "CLUSTERDOWN ") {
+		return true
+	}
+	if strings.HasPrefix(s, "TRYAGAIN ") {
+		return true
+	}
+
+	return false
+}
+
+func isRedisError(err error) bool {
+	_, ok := err.(proto.RedisError)
+	return ok
+}
+
+func isBadConn(err error, allowTimeout bool) bool {
+	if err == nil {
+		return false
+	}
+
+	if isRedisError(err) {
+		// Close connections in read only state in case domain addr is used
+		// and domain resolves to a different Redis Server. See #790.
+		return isReadOnlyError(err)
+	}
+
+	if allowTimeout {
+		if netErr, ok := err.(net.Error); ok && netErr.Timeout() {
+			return !netErr.Temporary()
+		}
+	}
+
+	return true
+}
+
+func isMovedError(err error) (moved bool, ask bool, addr string) {
+	if !isRedisError(err) {
+		return
+	}
+
+	s := err.Error()
+	switch {
+	case strings.HasPrefix(s, "MOVED "):
+		moved = true
+	case strings.HasPrefix(s, "ASK "):
+		ask = true
+	default:
+		return
+	}
+
+	ind := strings.LastIndex(s, " ")
+	if ind == -1 {
+		return false, false, ""
+	}
+	addr = s[ind+1:]
+	return
+}
+
+func isLoadingError(err error) bool {
+	return strings.HasPrefix(err.Error(), "LOADING ")
+}
+
+func isReadOnlyError(err error) bool {
+	return strings.HasPrefix(err.Error(), "READONLY ")
+}
+
+//------------------------------------------------------------------------------
+
+type timeoutError interface {
+	Timeout() bool
+}
diff --git a/vendor/github.com/go-redis/redis/v8/go.mod b/vendor/github.com/go-redis/redis/v8/go.mod
new file mode 100644
index 0000000..7be1b7f
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/go.mod
@@ -0,0 +1,11 @@
+module github.com/go-redis/redis/v8
+
+go 1.11
+
+require (
+	github.com/cespare/xxhash/v2 v2.1.1
+	github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f
+	github.com/onsi/ginkgo v1.14.2
+	github.com/onsi/gomega v1.10.3
+	go.opentelemetry.io/otel v0.13.0
+)
diff --git a/vendor/github.com/go-redis/redis/v8/go.sum b/vendor/github.com/go-redis/redis/v8/go.sum
new file mode 100644
index 0000000..e1e8ab1
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/go.sum
@@ -0,0 +1,82 @@
+github.com/cespare/xxhash/v2 v2.1.1 h1:6MnRN8NT7+YBpUIWxHtefFZOKTAPgGjpQSxqLNn0+qY=
+github.com/cespare/xxhash/v2 v2.1.1/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/davecgh/go-spew v1.1.0 h1:ZDRjVQ15GmhC3fiQ8ni8+OwkZQO4DARzQgrnXU1Liz8=
+github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/rVNCu3HqELle0jiPLLBs70cWOduZpkS1E78=
+github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
+github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
+github.com/fsnotify/fsnotify v1.4.9 h1:hsms1Qyu0jgnwNXIxa+/V/PDsU6CfLf6CNO8H7IWoS4=
+github.com/fsnotify/fsnotify v1.4.9/go.mod h1:znqG4EE+3YCdAaPaxE2ZRY/06pZUdp0tY4IgpuI1SZQ=
+github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
+github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
+github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
+github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
+github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
+github.com/golang/protobuf v1.4.2 h1:+Z5KGCizgyZCbGh1KZqA0fcLLkwbsjIzS4aV2v7wJX0=
+github.com/golang/protobuf v1.4.2/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.2 h1:X2ev0eStA3AbceY54o37/0PQ/UWqKEiiO2dKL5OPaFM=
+github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/hpcloud/tail v1.0.0/go.mod h1:ab1qPbhIpdTxEkNHXyeSf5vhxWSCs/tWer42PpOxQnU=
+github.com/nxadm/tail v1.4.4 h1:DQuhQpB1tVlglWS2hLQ5OV6B5r8aGxSrPc5Qo6uTN78=
+github.com/nxadm/tail v1.4.4/go.mod h1:kenIhsEOeOJmVchQTgglprH7qJGnHDVpk1VPCcaMI8A=
+github.com/onsi/ginkgo v1.6.0/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
+github.com/onsi/ginkgo v1.12.1/go.mod h1:zj2OWP4+oCPe1qIXoGWkgMRwljMUYCdkwsT2108oapk=
+github.com/onsi/ginkgo v1.14.2 h1:8mVmC9kjFFmA8H4pKMUhcblgifdkOIXPvbhN1T36q1M=
+github.com/onsi/ginkgo v1.14.2/go.mod h1:iSB4RoI2tjJc9BBv4NKIKWKya62Rps+oPG/Lv9klQyY=
+github.com/onsi/gomega v1.7.1/go.mod h1:XdKZgCCFLUoM/7CFJVPcG8C1xQ1AJ0vpAezJrB7JYyY=
+github.com/onsi/gomega v1.10.1/go.mod h1:iN09h71vgCQne3DLsj+A5owkum+a2tYe+TOCB1ybHNo=
+github.com/onsi/gomega v1.10.3 h1:gph6h/qe9GSUw1NhH1gp+qb+h8rXD8Cy60Z32Qw3ELA=
+github.com/onsi/gomega v1.10.3/go.mod h1:V9xEwhxec5O8UDM77eCW8vLymOMltsqPVYWrpDsH8xc=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.6.1 h1:hDPOHmpOpP40lSULcqw7IrRb/u7w6RpDC9399XyoNd0=
+github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
+go.opentelemetry.io/otel v0.13.0 h1:2isEnyzjjJZq6r2EKMsFj4TxiQiexsM04AVhwbR/oBA=
+go.opentelemetry.io/otel v0.13.0/go.mod h1:dlSNewoRYikTkotEnxdmuBHgzT+k/idJSfDv/FxEnOY=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/net v0.0.0-20180906233101-161cd47e91fd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20200520004742-59133d7f0dd7/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20201006153459-a7d1128ccaa0 h1:wBouT66WTYFXdxfVdz9sVWARVd/2vfGcmI45D2gj45M=
+golang.org/x/net v0.0.0-20201006153459-a7d1128ccaa0/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sys v0.0.0-20180909124046-d0be0721c37e/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190904154756-749cb33beabd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191005200804-aed5e4c7ecf9/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191120155948-bd437916bb0e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200323222414-85ca7c5b95cd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200519105757-fe76b779f299/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f h1:+Nyd8tzPX9R7BWHguqsrbFdRx3WQ/1ib8I44HXV5yTA=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
+golang.org/x/text v0.3.3 h1:cokOdA+Jmi5PJGXLlLllQSgYigAEfHXJAERHVMaCc2k=
+golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543 h1:E7g+9GITq07hpfrRu66IVDexMakfv52eLZ2CXBWiKr4=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
+google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
+google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
+google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
+google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
+google.golang.org/protobuf v1.23.0 h1:4MY060fB1DLGMB/7MBTLnwQUY6+F09GEiz6SsrNqyzM=
+google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/fsnotify.v1 v1.4.7/go.mod h1:Tz8NjZHkW78fSQdbUxIjBTcgA1z1m8ZHf0WmKUhAMys=
+gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 h1:uRGJdciOHaEIrze2W8Q3AKkepLTh2hOroT7a+7czfdQ=
+gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7/go.mod h1:dt/ZhP58zS4L8KSrWDmTeBkI65Dw0HsyUHuEVlX15mw=
+gopkg.in/yaml.v2 v2.2.4/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.3.0 h1:clyUAQHOM3G0M3f5vQj7LuJrETvjVot3Z5el9nffUtU=
+gopkg.in/yaml.v2 v2.3.0/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c h1:dUUwHk2QECo/6vqA44rthZ8ie2QXMNeKRTHCNY2nXvo=
+gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
diff --git a/vendor/github.com/go-redis/redis/v8/internal/arg.go b/vendor/github.com/go-redis/redis/v8/internal/arg.go
new file mode 100644
index 0000000..b97fa0d
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/arg.go
@@ -0,0 +1,56 @@
+package internal
+
+import (
+	"fmt"
+	"strconv"
+	"time"
+)
+
+func AppendArg(b []byte, v interface{}) []byte {
+	switch v := v.(type) {
+	case nil:
+		return append(b, "<nil>"...)
+	case string:
+		return appendUTF8String(b, Bytes(v))
+	case []byte:
+		return appendUTF8String(b, v)
+	case int:
+		return strconv.AppendInt(b, int64(v), 10)
+	case int8:
+		return strconv.AppendInt(b, int64(v), 10)
+	case int16:
+		return strconv.AppendInt(b, int64(v), 10)
+	case int32:
+		return strconv.AppendInt(b, int64(v), 10)
+	case int64:
+		return strconv.AppendInt(b, v, 10)
+	case uint:
+		return strconv.AppendUint(b, uint64(v), 10)
+	case uint8:
+		return strconv.AppendUint(b, uint64(v), 10)
+	case uint16:
+		return strconv.AppendUint(b, uint64(v), 10)
+	case uint32:
+		return strconv.AppendUint(b, uint64(v), 10)
+	case uint64:
+		return strconv.AppendUint(b, v, 10)
+	case float32:
+		return strconv.AppendFloat(b, float64(v), 'f', -1, 64)
+	case float64:
+		return strconv.AppendFloat(b, v, 'f', -1, 64)
+	case bool:
+		if v {
+			return append(b, "true"...)
+		}
+		return append(b, "false"...)
+	case time.Time:
+		return v.AppendFormat(b, time.RFC3339Nano)
+	default:
+		return append(b, fmt.Sprint(v)...)
+	}
+}
+
+func appendUTF8String(dst []byte, src []byte) []byte {
+	dst = append(dst, src...)
+	return dst
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/hashtag/hashtag.go b/vendor/github.com/go-redis/redis/v8/internal/hashtag/hashtag.go
new file mode 100644
index 0000000..2fc74ad
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/hashtag/hashtag.go
@@ -0,0 +1,78 @@
+package hashtag
+
+import (
+	"strings"
+
+	"github.com/go-redis/redis/v8/internal/rand"
+)
+
+const slotNumber = 16384
+
+// CRC16 implementation according to CCITT standards.
+// Copyright 2001-2010 Georges Menie (www.menie.org)
+// Copyright 2013 The Go Authors. All rights reserved.
+// http://redis.io/topics/cluster-spec#appendix-a-crc16-reference-implementation-in-ansi-c
+var crc16tab = [256]uint16{
+	0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50a5, 0x60c6, 0x70e7,
+	0x8108, 0x9129, 0xa14a, 0xb16b, 0xc18c, 0xd1ad, 0xe1ce, 0xf1ef,
+	0x1231, 0x0210, 0x3273, 0x2252, 0x52b5, 0x4294, 0x72f7, 0x62d6,
+	0x9339, 0x8318, 0xb37b, 0xa35a, 0xd3bd, 0xc39c, 0xf3ff, 0xe3de,
+	0x2462, 0x3443, 0x0420, 0x1401, 0x64e6, 0x74c7, 0x44a4, 0x5485,
+	0xa56a, 0xb54b, 0x8528, 0x9509, 0xe5ee, 0xf5cf, 0xc5ac, 0xd58d,
+	0x3653, 0x2672, 0x1611, 0x0630, 0x76d7, 0x66f6, 0x5695, 0x46b4,
+	0xb75b, 0xa77a, 0x9719, 0x8738, 0xf7df, 0xe7fe, 0xd79d, 0xc7bc,
+	0x48c4, 0x58e5, 0x6886, 0x78a7, 0x0840, 0x1861, 0x2802, 0x3823,
+	0xc9cc, 0xd9ed, 0xe98e, 0xf9af, 0x8948, 0x9969, 0xa90a, 0xb92b,
+	0x5af5, 0x4ad4, 0x7ab7, 0x6a96, 0x1a71, 0x0a50, 0x3a33, 0x2a12,
+	0xdbfd, 0xcbdc, 0xfbbf, 0xeb9e, 0x9b79, 0x8b58, 0xbb3b, 0xab1a,
+	0x6ca6, 0x7c87, 0x4ce4, 0x5cc5, 0x2c22, 0x3c03, 0x0c60, 0x1c41,
+	0xedae, 0xfd8f, 0xcdec, 0xddcd, 0xad2a, 0xbd0b, 0x8d68, 0x9d49,
+	0x7e97, 0x6eb6, 0x5ed5, 0x4ef4, 0x3e13, 0x2e32, 0x1e51, 0x0e70,
+	0xff9f, 0xefbe, 0xdfdd, 0xcffc, 0xbf1b, 0xaf3a, 0x9f59, 0x8f78,
+	0x9188, 0x81a9, 0xb1ca, 0xa1eb, 0xd10c, 0xc12d, 0xf14e, 0xe16f,
+	0x1080, 0x00a1, 0x30c2, 0x20e3, 0x5004, 0x4025, 0x7046, 0x6067,
+	0x83b9, 0x9398, 0xa3fb, 0xb3da, 0xc33d, 0xd31c, 0xe37f, 0xf35e,
+	0x02b1, 0x1290, 0x22f3, 0x32d2, 0x4235, 0x5214, 0x6277, 0x7256,
+	0xb5ea, 0xa5cb, 0x95a8, 0x8589, 0xf56e, 0xe54f, 0xd52c, 0xc50d,
+	0x34e2, 0x24c3, 0x14a0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
+	0xa7db, 0xb7fa, 0x8799, 0x97b8, 0xe75f, 0xf77e, 0xc71d, 0xd73c,
+	0x26d3, 0x36f2, 0x0691, 0x16b0, 0x6657, 0x7676, 0x4615, 0x5634,
+	0xd94c, 0xc96d, 0xf90e, 0xe92f, 0x99c8, 0x89e9, 0xb98a, 0xa9ab,
+	0x5844, 0x4865, 0x7806, 0x6827, 0x18c0, 0x08e1, 0x3882, 0x28a3,
+	0xcb7d, 0xdb5c, 0xeb3f, 0xfb1e, 0x8bf9, 0x9bd8, 0xabbb, 0xbb9a,
+	0x4a75, 0x5a54, 0x6a37, 0x7a16, 0x0af1, 0x1ad0, 0x2ab3, 0x3a92,
+	0xfd2e, 0xed0f, 0xdd6c, 0xcd4d, 0xbdaa, 0xad8b, 0x9de8, 0x8dc9,
+	0x7c26, 0x6c07, 0x5c64, 0x4c45, 0x3ca2, 0x2c83, 0x1ce0, 0x0cc1,
+	0xef1f, 0xff3e, 0xcf5d, 0xdf7c, 0xaf9b, 0xbfba, 0x8fd9, 0x9ff8,
+	0x6e17, 0x7e36, 0x4e55, 0x5e74, 0x2e93, 0x3eb2, 0x0ed1, 0x1ef0,
+}
+
+func Key(key string) string {
+	if s := strings.IndexByte(key, '{'); s > -1 {
+		if e := strings.IndexByte(key[s+1:], '}'); e > 0 {
+			return key[s+1 : s+e+1]
+		}
+	}
+	return key
+}
+
+func RandomSlot() int {
+	return rand.Intn(slotNumber)
+}
+
+// hashSlot returns a consistent slot number between 0 and 16383
+// for any given string key.
+func Slot(key string) int {
+	if key == "" {
+		return RandomSlot()
+	}
+	key = Key(key)
+	return int(crc16sum(key)) % slotNumber
+}
+
+func crc16sum(key string) (crc uint16) {
+	for i := 0; i < len(key); i++ {
+		crc = (crc << 8) ^ crc16tab[(byte(crc>>8)^key[i])&0x00ff]
+	}
+	return
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/instruments.go b/vendor/github.com/go-redis/redis/v8/internal/instruments.go
new file mode 100644
index 0000000..e837526
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/instruments.go
@@ -0,0 +1,33 @@
+package internal
+
+import (
+	"context"
+
+	"go.opentelemetry.io/otel/api/global"
+	"go.opentelemetry.io/otel/api/metric"
+)
+
+var (
+	// WritesCounter is a count of write commands performed.
+	WritesCounter metric.Int64Counter
+	// NewConnectionsCounter is a count of new connections.
+	NewConnectionsCounter metric.Int64Counter
+)
+
+func init() {
+	defer func() {
+		if r := recover(); r != nil {
+			Logger.Printf(context.Background(), "Error creating meter github.com/go-redis/redis for Instruments", r)
+		}
+	}()
+
+	meter := metric.Must(global.Meter("github.com/go-redis/redis"))
+
+	WritesCounter = meter.NewInt64Counter("redis.writes",
+		metric.WithDescription("the number of writes initiated"),
+	)
+
+	NewConnectionsCounter = meter.NewInt64Counter("redis.new_connections",
+		metric.WithDescription("the number of connections created"),
+	)
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/internal.go b/vendor/github.com/go-redis/redis/v8/internal/internal.go
new file mode 100644
index 0000000..4a59c59
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/internal.go
@@ -0,0 +1,29 @@
+package internal
+
+import (
+	"time"
+
+	"github.com/go-redis/redis/v8/internal/rand"
+)
+
+func RetryBackoff(retry int, minBackoff, maxBackoff time.Duration) time.Duration {
+	if retry < 0 {
+		panic("not reached")
+	}
+	if minBackoff == 0 {
+		return 0
+	}
+
+	d := minBackoff << uint(retry)
+	if d < minBackoff {
+		return maxBackoff
+	}
+
+	d = minBackoff + time.Duration(rand.Int63n(int64(d)))
+
+	if d > maxBackoff || d < minBackoff {
+		d = maxBackoff
+	}
+
+	return d
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/log.go b/vendor/github.com/go-redis/redis/v8/internal/log.go
new file mode 100644
index 0000000..3810f9e
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/log.go
@@ -0,0 +1,24 @@
+package internal
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"os"
+)
+
+type Logging interface {
+	Printf(ctx context.Context, format string, v ...interface{})
+}
+
+type logger struct {
+	log *log.Logger
+}
+
+func (l *logger) Printf(ctx context.Context, format string, v ...interface{}) {
+	_ = l.log.Output(2, fmt.Sprintf(format, v...))
+}
+
+var Logger Logging = &logger{
+	log: log.New(os.Stderr, "redis: ", log.LstdFlags|log.Lshortfile),
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/once.go b/vendor/github.com/go-redis/redis/v8/internal/once.go
new file mode 100644
index 0000000..64f4627
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/once.go
@@ -0,0 +1,60 @@
+/*
+Copyright 2014 The Camlistore Authors
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+     http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package internal
+
+import (
+	"sync"
+	"sync/atomic"
+)
+
+// A Once will perform a successful action exactly once.
+//
+// Unlike a sync.Once, this Once's func returns an error
+// and is re-armed on failure.
+type Once struct {
+	m    sync.Mutex
+	done uint32
+}
+
+// Do calls the function f if and only if Do has not been invoked
+// without error for this instance of Once.  In other words, given
+// 	var once Once
+// if once.Do(f) is called multiple times, only the first call will
+// invoke f, even if f has a different value in each invocation unless
+// f returns an error.  A new instance of Once is required for each
+// function to execute.
+//
+// Do is intended for initialization that must be run exactly once.  Since f
+// is niladic, it may be necessary to use a function literal to capture the
+// arguments to a function to be invoked by Do:
+// 	err := config.once.Do(func() error { return config.init(filename) })
+func (o *Once) Do(f func() error) error {
+	if atomic.LoadUint32(&o.done) == 1 {
+		return nil
+	}
+	// Slow-path.
+	o.m.Lock()
+	defer o.m.Unlock()
+	var err error
+	if o.done == 0 {
+		err = f()
+		if err == nil {
+			atomic.StoreUint32(&o.done, 1)
+		}
+	}
+	return err
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/pool/conn.go b/vendor/github.com/go-redis/redis/v8/internal/pool/conn.go
new file mode 100644
index 0000000..08a2071
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/pool/conn.go
@@ -0,0 +1,136 @@
+package pool
+
+import (
+	"bufio"
+	"context"
+	"net"
+	"sync/atomic"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/proto"
+	"go.opentelemetry.io/otel/api/trace"
+)
+
+var noDeadline = time.Time{}
+
+type Conn struct {
+	usedAt  int64 // atomic
+	netConn net.Conn
+
+	rd *proto.Reader
+	bw *bufio.Writer
+	wr *proto.Writer
+
+	Inited    bool
+	pooled    bool
+	createdAt time.Time
+}
+
+func NewConn(netConn net.Conn) *Conn {
+	cn := &Conn{
+		netConn:   netConn,
+		createdAt: time.Now(),
+	}
+	cn.rd = proto.NewReader(netConn)
+	cn.bw = bufio.NewWriter(netConn)
+	cn.wr = proto.NewWriter(cn.bw)
+	cn.SetUsedAt(time.Now())
+	return cn
+}
+
+func (cn *Conn) UsedAt() time.Time {
+	unix := atomic.LoadInt64(&cn.usedAt)
+	return time.Unix(unix, 0)
+}
+
+func (cn *Conn) SetUsedAt(tm time.Time) {
+	atomic.StoreInt64(&cn.usedAt, tm.Unix())
+}
+
+func (cn *Conn) SetNetConn(netConn net.Conn) {
+	cn.netConn = netConn
+	cn.rd.Reset(netConn)
+	cn.bw.Reset(netConn)
+}
+
+func (cn *Conn) Write(b []byte) (int, error) {
+	return cn.netConn.Write(b)
+}
+
+func (cn *Conn) RemoteAddr() net.Addr {
+	if cn.netConn != nil {
+		return cn.netConn.RemoteAddr()
+	}
+	return nil
+}
+
+func (cn *Conn) WithReader(ctx context.Context, timeout time.Duration, fn func(rd *proto.Reader) error) error {
+	return internal.WithSpan(ctx, "redis.with_reader", func(ctx context.Context, span trace.Span) error {
+		if err := cn.netConn.SetReadDeadline(cn.deadline(ctx, timeout)); err != nil {
+			return internal.RecordError(ctx, err)
+		}
+		if err := fn(cn.rd); err != nil {
+			return internal.RecordError(ctx, err)
+		}
+		return nil
+	})
+}
+
+func (cn *Conn) WithWriter(
+	ctx context.Context, timeout time.Duration, fn func(wr *proto.Writer) error,
+) error {
+	return internal.WithSpan(ctx, "redis.with_writer", func(ctx context.Context, span trace.Span) error {
+		if err := cn.netConn.SetWriteDeadline(cn.deadline(ctx, timeout)); err != nil {
+			return internal.RecordError(ctx, err)
+		}
+
+		if cn.bw.Buffered() > 0 {
+			cn.bw.Reset(cn.netConn)
+		}
+
+		if err := fn(cn.wr); err != nil {
+			return internal.RecordError(ctx, err)
+		}
+
+		if err := cn.bw.Flush(); err != nil {
+			return internal.RecordError(ctx, err)
+		}
+
+		internal.WritesCounter.Add(ctx, 1)
+
+		return nil
+	})
+}
+
+func (cn *Conn) Close() error {
+	return cn.netConn.Close()
+}
+
+func (cn *Conn) deadline(ctx context.Context, timeout time.Duration) time.Time {
+	tm := time.Now()
+	cn.SetUsedAt(tm)
+
+	if timeout > 0 {
+		tm = tm.Add(timeout)
+	}
+
+	if ctx != nil {
+		deadline, ok := ctx.Deadline()
+		if ok {
+			if timeout == 0 {
+				return deadline
+			}
+			if deadline.Before(tm) {
+				return deadline
+			}
+			return tm
+		}
+	}
+
+	if timeout > 0 {
+		return tm
+	}
+
+	return noDeadline
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/pool/pool.go b/vendor/github.com/go-redis/redis/v8/internal/pool/pool.go
new file mode 100644
index 0000000..355742b
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/pool/pool.go
@@ -0,0 +1,524 @@
+package pool
+
+import (
+	"context"
+	"errors"
+	"net"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+)
+
+var (
+	ErrClosed      = errors.New("redis: client is closed")
+	ErrPoolTimeout = errors.New("redis: connection pool timeout")
+)
+
+var timers = sync.Pool{
+	New: func() interface{} {
+		t := time.NewTimer(time.Hour)
+		t.Stop()
+		return t
+	},
+}
+
+// Stats contains pool state information and accumulated stats.
+type Stats struct {
+	Hits     uint32 // number of times free connection was found in the pool
+	Misses   uint32 // number of times free connection was NOT found in the pool
+	Timeouts uint32 // number of times a wait timeout occurred
+
+	TotalConns uint32 // number of total connections in the pool
+	IdleConns  uint32 // number of idle connections in the pool
+	StaleConns uint32 // number of stale connections removed from the pool
+}
+
+type Pooler interface {
+	NewConn(context.Context) (*Conn, error)
+	CloseConn(*Conn) error
+
+	Get(context.Context) (*Conn, error)
+	Put(context.Context, *Conn)
+	Remove(context.Context, *Conn, error)
+
+	Len() int
+	IdleLen() int
+	Stats() *Stats
+
+	Close() error
+}
+
+type Options struct {
+	Dialer  func(context.Context) (net.Conn, error)
+	OnClose func(*Conn) error
+
+	PoolSize           int
+	MinIdleConns       int
+	MaxConnAge         time.Duration
+	PoolTimeout        time.Duration
+	IdleTimeout        time.Duration
+	IdleCheckFrequency time.Duration
+}
+
+type lastDialErrorWrap struct {
+	err error
+}
+
+type ConnPool struct {
+	opt *Options
+
+	dialErrorsNum uint32 // atomic
+
+	lastDialError atomic.Value
+
+	queue chan struct{}
+
+	connsMu      sync.Mutex
+	conns        []*Conn
+	idleConns    []*Conn
+	poolSize     int
+	idleConnsLen int
+
+	stats Stats
+
+	_closed  uint32 // atomic
+	closedCh chan struct{}
+}
+
+var _ Pooler = (*ConnPool)(nil)
+
+func NewConnPool(opt *Options) *ConnPool {
+	p := &ConnPool{
+		opt: opt,
+
+		queue:     make(chan struct{}, opt.PoolSize),
+		conns:     make([]*Conn, 0, opt.PoolSize),
+		idleConns: make([]*Conn, 0, opt.PoolSize),
+		closedCh:  make(chan struct{}),
+	}
+
+	p.connsMu.Lock()
+	p.checkMinIdleConns()
+	p.connsMu.Unlock()
+
+	if opt.IdleTimeout > 0 && opt.IdleCheckFrequency > 0 {
+		go p.reaper(opt.IdleCheckFrequency)
+	}
+
+	return p
+}
+
+func (p *ConnPool) checkMinIdleConns() {
+	if p.opt.MinIdleConns == 0 {
+		return
+	}
+	for p.poolSize < p.opt.PoolSize && p.idleConnsLen < p.opt.MinIdleConns {
+		p.poolSize++
+		p.idleConnsLen++
+		go func() {
+			err := p.addIdleConn()
+			if err != nil {
+				p.connsMu.Lock()
+				p.poolSize--
+				p.idleConnsLen--
+				p.connsMu.Unlock()
+			}
+		}()
+	}
+}
+
+func (p *ConnPool) addIdleConn() error {
+	cn, err := p.dialConn(context.TODO(), true)
+	if err != nil {
+		return err
+	}
+
+	p.connsMu.Lock()
+	p.conns = append(p.conns, cn)
+	p.idleConns = append(p.idleConns, cn)
+	p.connsMu.Unlock()
+	return nil
+}
+
+func (p *ConnPool) NewConn(ctx context.Context) (*Conn, error) {
+	return p.newConn(ctx, false)
+}
+
+func (p *ConnPool) newConn(ctx context.Context, pooled bool) (*Conn, error) {
+	cn, err := p.dialConn(ctx, pooled)
+	if err != nil {
+		return nil, err
+	}
+
+	p.connsMu.Lock()
+	p.conns = append(p.conns, cn)
+	if pooled {
+		// If pool is full remove the cn on next Put.
+		if p.poolSize >= p.opt.PoolSize {
+			cn.pooled = false
+		} else {
+			p.poolSize++
+		}
+	}
+	p.connsMu.Unlock()
+
+	return cn, nil
+}
+
+func (p *ConnPool) dialConn(ctx context.Context, pooled bool) (*Conn, error) {
+	if p.closed() {
+		return nil, ErrClosed
+	}
+
+	if atomic.LoadUint32(&p.dialErrorsNum) >= uint32(p.opt.PoolSize) {
+		return nil, p.getLastDialError()
+	}
+
+	netConn, err := p.opt.Dialer(ctx)
+	if err != nil {
+		p.setLastDialError(err)
+		if atomic.AddUint32(&p.dialErrorsNum, 1) == uint32(p.opt.PoolSize) {
+			go p.tryDial()
+		}
+		return nil, err
+	}
+
+	internal.NewConnectionsCounter.Add(ctx, 1)
+	cn := NewConn(netConn)
+	cn.pooled = pooled
+	return cn, nil
+}
+
+func (p *ConnPool) tryDial() {
+	for {
+		if p.closed() {
+			return
+		}
+
+		conn, err := p.opt.Dialer(context.Background())
+		if err != nil {
+			p.setLastDialError(err)
+			time.Sleep(time.Second)
+			continue
+		}
+
+		atomic.StoreUint32(&p.dialErrorsNum, 0)
+		_ = conn.Close()
+		return
+	}
+}
+
+func (p *ConnPool) setLastDialError(err error) {
+	p.lastDialError.Store(&lastDialErrorWrap{err: err})
+}
+
+func (p *ConnPool) getLastDialError() error {
+	err, _ := p.lastDialError.Load().(*lastDialErrorWrap)
+	if err != nil {
+		return err.err
+	}
+	return nil
+}
+
+// Get returns existed connection from the pool or creates a new one.
+func (p *ConnPool) Get(ctx context.Context) (*Conn, error) {
+	if p.closed() {
+		return nil, ErrClosed
+	}
+
+	err := p.waitTurn(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	for {
+		p.connsMu.Lock()
+		cn := p.popIdle()
+		p.connsMu.Unlock()
+
+		if cn == nil {
+			break
+		}
+
+		if p.isStaleConn(cn) {
+			_ = p.CloseConn(cn)
+			continue
+		}
+
+		atomic.AddUint32(&p.stats.Hits, 1)
+		return cn, nil
+	}
+
+	atomic.AddUint32(&p.stats.Misses, 1)
+
+	newcn, err := p.newConn(ctx, true)
+	if err != nil {
+		p.freeTurn()
+		return nil, err
+	}
+
+	return newcn, nil
+}
+
+func (p *ConnPool) getTurn() {
+	p.queue <- struct{}{}
+}
+
+func (p *ConnPool) waitTurn(ctx context.Context) error {
+	select {
+	case <-ctx.Done():
+		return ctx.Err()
+	default:
+	}
+
+	select {
+	case p.queue <- struct{}{}:
+		return nil
+	default:
+	}
+
+	timer := timers.Get().(*time.Timer)
+	timer.Reset(p.opt.PoolTimeout)
+
+	select {
+	case <-ctx.Done():
+		if !timer.Stop() {
+			<-timer.C
+		}
+		timers.Put(timer)
+		return ctx.Err()
+	case p.queue <- struct{}{}:
+		if !timer.Stop() {
+			<-timer.C
+		}
+		timers.Put(timer)
+		return nil
+	case <-timer.C:
+		timers.Put(timer)
+		atomic.AddUint32(&p.stats.Timeouts, 1)
+		return ErrPoolTimeout
+	}
+}
+
+func (p *ConnPool) freeTurn() {
+	<-p.queue
+}
+
+func (p *ConnPool) popIdle() *Conn {
+	if len(p.idleConns) == 0 {
+		return nil
+	}
+
+	idx := len(p.idleConns) - 1
+	cn := p.idleConns[idx]
+	p.idleConns = p.idleConns[:idx]
+	p.idleConnsLen--
+	p.checkMinIdleConns()
+	return cn
+}
+
+func (p *ConnPool) Put(ctx context.Context, cn *Conn) {
+	if cn.rd.Buffered() > 0 {
+		internal.Logger.Printf(ctx, "Conn has unread data")
+		p.Remove(ctx, cn, BadConnError{})
+		return
+	}
+
+	if !cn.pooled {
+		p.Remove(ctx, cn, nil)
+		return
+	}
+
+	p.connsMu.Lock()
+	p.idleConns = append(p.idleConns, cn)
+	p.idleConnsLen++
+	p.connsMu.Unlock()
+	p.freeTurn()
+}
+
+func (p *ConnPool) Remove(ctx context.Context, cn *Conn, reason error) {
+	p.removeConnWithLock(cn)
+	p.freeTurn()
+	_ = p.closeConn(cn)
+}
+
+func (p *ConnPool) CloseConn(cn *Conn) error {
+	p.removeConnWithLock(cn)
+	return p.closeConn(cn)
+}
+
+func (p *ConnPool) removeConnWithLock(cn *Conn) {
+	p.connsMu.Lock()
+	p.removeConn(cn)
+	p.connsMu.Unlock()
+}
+
+func (p *ConnPool) removeConn(cn *Conn) {
+	for i, c := range p.conns {
+		if c == cn {
+			p.conns = append(p.conns[:i], p.conns[i+1:]...)
+			if cn.pooled {
+				p.poolSize--
+				p.checkMinIdleConns()
+			}
+			return
+		}
+	}
+}
+
+func (p *ConnPool) closeConn(cn *Conn) error {
+	if p.opt.OnClose != nil {
+		_ = p.opt.OnClose(cn)
+	}
+	return cn.Close()
+}
+
+// Len returns total number of connections.
+func (p *ConnPool) Len() int {
+	p.connsMu.Lock()
+	n := len(p.conns)
+	p.connsMu.Unlock()
+	return n
+}
+
+// IdleLen returns number of idle connections.
+func (p *ConnPool) IdleLen() int {
+	p.connsMu.Lock()
+	n := p.idleConnsLen
+	p.connsMu.Unlock()
+	return n
+}
+
+func (p *ConnPool) Stats() *Stats {
+	idleLen := p.IdleLen()
+	return &Stats{
+		Hits:     atomic.LoadUint32(&p.stats.Hits),
+		Misses:   atomic.LoadUint32(&p.stats.Misses),
+		Timeouts: atomic.LoadUint32(&p.stats.Timeouts),
+
+		TotalConns: uint32(p.Len()),
+		IdleConns:  uint32(idleLen),
+		StaleConns: atomic.LoadUint32(&p.stats.StaleConns),
+	}
+}
+
+func (p *ConnPool) closed() bool {
+	return atomic.LoadUint32(&p._closed) == 1
+}
+
+func (p *ConnPool) Filter(fn func(*Conn) bool) error {
+	p.connsMu.Lock()
+	defer p.connsMu.Unlock()
+
+	var firstErr error
+	for _, cn := range p.conns {
+		if fn(cn) {
+			if err := p.closeConn(cn); err != nil && firstErr == nil {
+				firstErr = err
+			}
+		}
+	}
+	return firstErr
+}
+
+func (p *ConnPool) Close() error {
+	if !atomic.CompareAndSwapUint32(&p._closed, 0, 1) {
+		return ErrClosed
+	}
+	close(p.closedCh)
+
+	var firstErr error
+	p.connsMu.Lock()
+	for _, cn := range p.conns {
+		if err := p.closeConn(cn); err != nil && firstErr == nil {
+			firstErr = err
+		}
+	}
+	p.conns = nil
+	p.poolSize = 0
+	p.idleConns = nil
+	p.idleConnsLen = 0
+	p.connsMu.Unlock()
+
+	return firstErr
+}
+
+func (p *ConnPool) reaper(frequency time.Duration) {
+	ticker := time.NewTicker(frequency)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ticker.C:
+			// It is possible that ticker and closedCh arrive together,
+			// and select pseudo-randomly pick ticker case, we double
+			// check here to prevent being executed after closed.
+			if p.closed() {
+				return
+			}
+			_, err := p.ReapStaleConns()
+			if err != nil {
+				internal.Logger.Printf(context.Background(), "ReapStaleConns failed: %s", err)
+				continue
+			}
+		case <-p.closedCh:
+			return
+		}
+	}
+}
+
+func (p *ConnPool) ReapStaleConns() (int, error) {
+	var n int
+	for {
+		p.getTurn()
+
+		p.connsMu.Lock()
+		cn := p.reapStaleConn()
+		p.connsMu.Unlock()
+		p.freeTurn()
+
+		if cn != nil {
+			_ = p.closeConn(cn)
+			n++
+		} else {
+			break
+		}
+	}
+	atomic.AddUint32(&p.stats.StaleConns, uint32(n))
+	return n, nil
+}
+
+func (p *ConnPool) reapStaleConn() *Conn {
+	if len(p.idleConns) == 0 {
+		return nil
+	}
+
+	cn := p.idleConns[0]
+	if !p.isStaleConn(cn) {
+		return nil
+	}
+
+	p.idleConns = append(p.idleConns[:0], p.idleConns[1:]...)
+	p.idleConnsLen--
+	p.removeConn(cn)
+
+	return cn
+}
+
+func (p *ConnPool) isStaleConn(cn *Conn) bool {
+	if p.opt.IdleTimeout == 0 && p.opt.MaxConnAge == 0 {
+		return false
+	}
+
+	now := time.Now()
+	if p.opt.IdleTimeout > 0 && now.Sub(cn.UsedAt()) >= p.opt.IdleTimeout {
+		return true
+	}
+	if p.opt.MaxConnAge > 0 && now.Sub(cn.createdAt) >= p.opt.MaxConnAge {
+		return true
+	}
+
+	return false
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/pool/pool_single.go b/vendor/github.com/go-redis/redis/v8/internal/pool/pool_single.go
new file mode 100644
index 0000000..5a3fde1
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/pool/pool_single.go
@@ -0,0 +1,58 @@
+package pool
+
+import "context"
+
+type SingleConnPool struct {
+	pool      Pooler
+	cn        *Conn
+	stickyErr error
+}
+
+var _ Pooler = (*SingleConnPool)(nil)
+
+func NewSingleConnPool(pool Pooler, cn *Conn) *SingleConnPool {
+	return &SingleConnPool{
+		pool: pool,
+		cn:   cn,
+	}
+}
+
+func (p *SingleConnPool) NewConn(ctx context.Context) (*Conn, error) {
+	return p.pool.NewConn(ctx)
+}
+
+func (p *SingleConnPool) CloseConn(cn *Conn) error {
+	return p.pool.CloseConn(cn)
+}
+
+func (p *SingleConnPool) Get(ctx context.Context) (*Conn, error) {
+	if p.stickyErr != nil {
+		return nil, p.stickyErr
+	}
+	return p.cn, nil
+}
+
+func (p *SingleConnPool) Put(ctx context.Context, cn *Conn) {}
+
+func (p *SingleConnPool) Remove(ctx context.Context, cn *Conn, reason error) {
+	p.cn = nil
+	p.stickyErr = reason
+}
+
+func (p *SingleConnPool) Close() error {
+	p.cn = nil
+	p.stickyErr = ErrClosed
+	return nil
+}
+
+func (p *SingleConnPool) Len() int {
+	return 0
+}
+
+func (p *SingleConnPool) IdleLen() int {
+	return 0
+}
+
+func (p *SingleConnPool) Stats() *Stats {
+	return &Stats{}
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/pool/pool_sticky.go b/vendor/github.com/go-redis/redis/v8/internal/pool/pool_sticky.go
new file mode 100644
index 0000000..c3e7e7c
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/pool/pool_sticky.go
@@ -0,0 +1,202 @@
+package pool
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"sync/atomic"
+)
+
+const (
+	stateDefault = 0
+	stateInited  = 1
+	stateClosed  = 2
+)
+
+type BadConnError struct {
+	wrapped error
+}
+
+var _ error = (*BadConnError)(nil)
+
+func (e BadConnError) Error() string {
+	s := "redis: Conn is in a bad state"
+	if e.wrapped != nil {
+		s += ": " + e.wrapped.Error()
+	}
+	return s
+}
+
+func (e BadConnError) Unwrap() error {
+	return e.wrapped
+}
+
+//------------------------------------------------------------------------------
+
+type StickyConnPool struct {
+	pool   Pooler
+	shared int32 // atomic
+
+	state uint32 // atomic
+	ch    chan *Conn
+
+	_badConnError atomic.Value
+}
+
+var _ Pooler = (*StickyConnPool)(nil)
+
+func NewStickyConnPool(pool Pooler) *StickyConnPool {
+	p, ok := pool.(*StickyConnPool)
+	if !ok {
+		p = &StickyConnPool{
+			pool: pool,
+			ch:   make(chan *Conn, 1),
+		}
+	}
+	atomic.AddInt32(&p.shared, 1)
+	return p
+}
+
+func (p *StickyConnPool) NewConn(ctx context.Context) (*Conn, error) {
+	return p.pool.NewConn(ctx)
+}
+
+func (p *StickyConnPool) CloseConn(cn *Conn) error {
+	return p.pool.CloseConn(cn)
+}
+
+func (p *StickyConnPool) Get(ctx context.Context) (*Conn, error) {
+	// In worst case this races with Close which is not a very common operation.
+	for i := 0; i < 1000; i++ {
+		switch atomic.LoadUint32(&p.state) {
+		case stateDefault:
+			cn, err := p.pool.Get(ctx)
+			if err != nil {
+				return nil, err
+			}
+			if atomic.CompareAndSwapUint32(&p.state, stateDefault, stateInited) {
+				return cn, nil
+			}
+			p.pool.Remove(ctx, cn, ErrClosed)
+		case stateInited:
+			if err := p.badConnError(); err != nil {
+				return nil, err
+			}
+			cn, ok := <-p.ch
+			if !ok {
+				return nil, ErrClosed
+			}
+			return cn, nil
+		case stateClosed:
+			return nil, ErrClosed
+		default:
+			panic("not reached")
+		}
+	}
+	return nil, fmt.Errorf("redis: StickyConnPool.Get: infinite loop")
+}
+
+func (p *StickyConnPool) Put(ctx context.Context, cn *Conn) {
+	defer func() {
+		if recover() != nil {
+			p.freeConn(ctx, cn)
+		}
+	}()
+	p.ch <- cn
+}
+
+func (p *StickyConnPool) freeConn(ctx context.Context, cn *Conn) {
+	if err := p.badConnError(); err != nil {
+		p.pool.Remove(ctx, cn, err)
+	} else {
+		p.pool.Put(ctx, cn)
+	}
+}
+
+func (p *StickyConnPool) Remove(ctx context.Context, cn *Conn, reason error) {
+	defer func() {
+		if recover() != nil {
+			p.pool.Remove(ctx, cn, ErrClosed)
+		}
+	}()
+	p._badConnError.Store(BadConnError{wrapped: reason})
+	p.ch <- cn
+}
+
+func (p *StickyConnPool) Close() error {
+	if shared := atomic.AddInt32(&p.shared, -1); shared > 0 {
+		return nil
+	}
+
+	for i := 0; i < 1000; i++ {
+		state := atomic.LoadUint32(&p.state)
+		if state == stateClosed {
+			return ErrClosed
+		}
+		if atomic.CompareAndSwapUint32(&p.state, state, stateClosed) {
+			close(p.ch)
+			cn, ok := <-p.ch
+			if ok {
+				p.freeConn(context.TODO(), cn)
+			}
+			return nil
+		}
+	}
+
+	return errors.New("redis: StickyConnPool.Close: infinite loop")
+}
+
+func (p *StickyConnPool) Reset(ctx context.Context) error {
+	if p.badConnError() == nil {
+		return nil
+	}
+
+	select {
+	case cn, ok := <-p.ch:
+		if !ok {
+			return ErrClosed
+		}
+		p.pool.Remove(ctx, cn, ErrClosed)
+		p._badConnError.Store(BadConnError{wrapped: nil})
+	default:
+		return errors.New("redis: StickyConnPool does not have a Conn")
+	}
+
+	if !atomic.CompareAndSwapUint32(&p.state, stateInited, stateDefault) {
+		state := atomic.LoadUint32(&p.state)
+		return fmt.Errorf("redis: invalid StickyConnPool state: %d", state)
+	}
+
+	return nil
+}
+
+func (p *StickyConnPool) badConnError() error {
+	if v := p._badConnError.Load(); v != nil {
+		err := v.(BadConnError)
+		if err.wrapped != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func (p *StickyConnPool) Len() int {
+	switch atomic.LoadUint32(&p.state) {
+	case stateDefault:
+		return 0
+	case stateInited:
+		return 1
+	case stateClosed:
+		return 0
+	default:
+		panic("not reached")
+	}
+}
+
+func (p *StickyConnPool) IdleLen() int {
+	return len(p.ch)
+}
+
+func (p *StickyConnPool) Stats() *Stats {
+	return &Stats{}
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/proto/reader.go b/vendor/github.com/go-redis/redis/v8/internal/proto/reader.go
new file mode 100644
index 0000000..0fbc51e
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/proto/reader.go
@@ -0,0 +1,331 @@
+package proto
+
+import (
+	"bufio"
+	"fmt"
+	"io"
+
+	"github.com/go-redis/redis/v8/internal/util"
+)
+
+const (
+	ErrorReply  = '-'
+	StatusReply = '+'
+	IntReply    = ':'
+	StringReply = '$'
+	ArrayReply  = '*'
+)
+
+//------------------------------------------------------------------------------
+
+const Nil = RedisError("redis: nil")
+
+type RedisError string
+
+func (e RedisError) Error() string { return string(e) }
+
+func (RedisError) RedisError() {}
+
+//------------------------------------------------------------------------------
+
+type MultiBulkParse func(*Reader, int64) (interface{}, error)
+
+type Reader struct {
+	rd   *bufio.Reader
+	_buf []byte
+}
+
+func NewReader(rd io.Reader) *Reader {
+	return &Reader{
+		rd:   bufio.NewReader(rd),
+		_buf: make([]byte, 64),
+	}
+}
+
+func (r *Reader) Buffered() int {
+	return r.rd.Buffered()
+}
+
+func (r *Reader) Peek(n int) ([]byte, error) {
+	return r.rd.Peek(n)
+}
+
+func (r *Reader) Reset(rd io.Reader) {
+	r.rd.Reset(rd)
+}
+
+func (r *Reader) ReadLine() ([]byte, error) {
+	line, err := r.readLine()
+	if err != nil {
+		return nil, err
+	}
+	if isNilReply(line) {
+		return nil, Nil
+	}
+	return line, nil
+}
+
+// readLine that returns an error if:
+//   - there is a pending read error;
+//   - or line does not end with \r\n.
+func (r *Reader) readLine() ([]byte, error) {
+	b, err := r.rd.ReadSlice('\n')
+	if err != nil {
+		if err != bufio.ErrBufferFull {
+			return nil, err
+		}
+
+		full := make([]byte, len(b))
+		copy(full, b)
+
+		b, err = r.rd.ReadBytes('\n')
+		if err != nil {
+			return nil, err
+		}
+
+		full = append(full, b...)
+		b = full
+	}
+	if len(b) <= 2 || b[len(b)-1] != '\n' || b[len(b)-2] != '\r' {
+		return nil, fmt.Errorf("redis: invalid reply: %q", b)
+	}
+	return b[:len(b)-2], nil
+}
+
+func (r *Reader) ReadReply(m MultiBulkParse) (interface{}, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return nil, err
+	}
+
+	switch line[0] {
+	case ErrorReply:
+		return nil, ParseErrorReply(line)
+	case StatusReply:
+		return string(line[1:]), nil
+	case IntReply:
+		return util.ParseInt(line[1:], 10, 64)
+	case StringReply:
+		return r.readStringReply(line)
+	case ArrayReply:
+		n, err := parseArrayLen(line)
+		if err != nil {
+			return nil, err
+		}
+		if m == nil {
+			err := fmt.Errorf("redis: got %.100q, but multi bulk parser is nil", line)
+			return nil, err
+		}
+		return m(r, n)
+	}
+	return nil, fmt.Errorf("redis: can't parse %.100q", line)
+}
+
+func (r *Reader) ReadIntReply() (int64, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return 0, err
+	}
+	switch line[0] {
+	case ErrorReply:
+		return 0, ParseErrorReply(line)
+	case IntReply:
+		return util.ParseInt(line[1:], 10, 64)
+	default:
+		return 0, fmt.Errorf("redis: can't parse int reply: %.100q", line)
+	}
+}
+
+func (r *Reader) ReadString() (string, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return "", err
+	}
+	switch line[0] {
+	case ErrorReply:
+		return "", ParseErrorReply(line)
+	case StringReply:
+		return r.readStringReply(line)
+	case StatusReply:
+		return string(line[1:]), nil
+	case IntReply:
+		return string(line[1:]), nil
+	default:
+		return "", fmt.Errorf("redis: can't parse reply=%.100q reading string", line)
+	}
+}
+
+func (r *Reader) readStringReply(line []byte) (string, error) {
+	if isNilReply(line) {
+		return "", Nil
+	}
+
+	replyLen, err := util.Atoi(line[1:])
+	if err != nil {
+		return "", err
+	}
+
+	b := make([]byte, replyLen+2)
+	_, err = io.ReadFull(r.rd, b)
+	if err != nil {
+		return "", err
+	}
+
+	return util.BytesToString(b[:replyLen]), nil
+}
+
+func (r *Reader) ReadArrayReply(m MultiBulkParse) (interface{}, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return nil, err
+	}
+	switch line[0] {
+	case ErrorReply:
+		return nil, ParseErrorReply(line)
+	case ArrayReply:
+		n, err := parseArrayLen(line)
+		if err != nil {
+			return nil, err
+		}
+		return m(r, n)
+	default:
+		return nil, fmt.Errorf("redis: can't parse array reply: %.100q", line)
+	}
+}
+
+func (r *Reader) ReadArrayLen() (int, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return 0, err
+	}
+	switch line[0] {
+	case ErrorReply:
+		return 0, ParseErrorReply(line)
+	case ArrayReply:
+		n, err := parseArrayLen(line)
+		if err != nil {
+			return 0, err
+		}
+		return int(n), nil
+	default:
+		return 0, fmt.Errorf("redis: can't parse array reply: %.100q", line)
+	}
+}
+
+func (r *Reader) ReadScanReply() ([]string, uint64, error) {
+	n, err := r.ReadArrayLen()
+	if err != nil {
+		return nil, 0, err
+	}
+	if n != 2 {
+		return nil, 0, fmt.Errorf("redis: got %d elements in scan reply, expected 2", n)
+	}
+
+	cursor, err := r.ReadUint()
+	if err != nil {
+		return nil, 0, err
+	}
+
+	n, err = r.ReadArrayLen()
+	if err != nil {
+		return nil, 0, err
+	}
+
+	keys := make([]string, n)
+
+	for i := 0; i < n; i++ {
+		key, err := r.ReadString()
+		if err != nil {
+			return nil, 0, err
+		}
+		keys[i] = key
+	}
+
+	return keys, cursor, err
+}
+
+func (r *Reader) ReadInt() (int64, error) {
+	b, err := r.readTmpBytesReply()
+	if err != nil {
+		return 0, err
+	}
+	return util.ParseInt(b, 10, 64)
+}
+
+func (r *Reader) ReadUint() (uint64, error) {
+	b, err := r.readTmpBytesReply()
+	if err != nil {
+		return 0, err
+	}
+	return util.ParseUint(b, 10, 64)
+}
+
+func (r *Reader) ReadFloatReply() (float64, error) {
+	b, err := r.readTmpBytesReply()
+	if err != nil {
+		return 0, err
+	}
+	return util.ParseFloat(b, 64)
+}
+
+func (r *Reader) readTmpBytesReply() ([]byte, error) {
+	line, err := r.ReadLine()
+	if err != nil {
+		return nil, err
+	}
+	switch line[0] {
+	case ErrorReply:
+		return nil, ParseErrorReply(line)
+	case StringReply:
+		return r._readTmpBytesReply(line)
+	case StatusReply:
+		return line[1:], nil
+	default:
+		return nil, fmt.Errorf("redis: can't parse string reply: %.100q", line)
+	}
+}
+
+func (r *Reader) _readTmpBytesReply(line []byte) ([]byte, error) {
+	if isNilReply(line) {
+		return nil, Nil
+	}
+
+	replyLen, err := util.Atoi(line[1:])
+	if err != nil {
+		return nil, err
+	}
+
+	buf := r.buf(replyLen + 2)
+	_, err = io.ReadFull(r.rd, buf)
+	if err != nil {
+		return nil, err
+	}
+
+	return buf[:replyLen], nil
+}
+
+func (r *Reader) buf(n int) []byte {
+	if n <= cap(r._buf) {
+		return r._buf[:n]
+	}
+	d := n - cap(r._buf)
+	r._buf = append(r._buf, make([]byte, d)...)
+	return r._buf
+}
+
+func isNilReply(b []byte) bool {
+	return len(b) == 3 &&
+		(b[0] == StringReply || b[0] == ArrayReply) &&
+		b[1] == '-' && b[2] == '1'
+}
+
+func ParseErrorReply(line []byte) error {
+	return RedisError(string(line[1:]))
+}
+
+func parseArrayLen(line []byte) (int64, error) {
+	if isNilReply(line) {
+		return 0, Nil
+	}
+	return util.ParseInt(line[1:], 10, 64)
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/proto/scan.go b/vendor/github.com/go-redis/redis/v8/internal/proto/scan.go
new file mode 100644
index 0000000..08d18d3
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/proto/scan.go
@@ -0,0 +1,173 @@
+package proto
+
+import (
+	"encoding"
+	"fmt"
+	"reflect"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal/util"
+)
+
+// Scan parses bytes `b` to `v` with appropriate type.
+// nolint: gocyclo
+func Scan(b []byte, v interface{}) error {
+	switch v := v.(type) {
+	case nil:
+		return fmt.Errorf("redis: Scan(nil)")
+	case *string:
+		*v = util.BytesToString(b)
+		return nil
+	case *[]byte:
+		*v = b
+		return nil
+	case *int:
+		var err error
+		*v, err = util.Atoi(b)
+		return err
+	case *int8:
+		n, err := util.ParseInt(b, 10, 8)
+		if err != nil {
+			return err
+		}
+		*v = int8(n)
+		return nil
+	case *int16:
+		n, err := util.ParseInt(b, 10, 16)
+		if err != nil {
+			return err
+		}
+		*v = int16(n)
+		return nil
+	case *int32:
+		n, err := util.ParseInt(b, 10, 32)
+		if err != nil {
+			return err
+		}
+		*v = int32(n)
+		return nil
+	case *int64:
+		n, err := util.ParseInt(b, 10, 64)
+		if err != nil {
+			return err
+		}
+		*v = n
+		return nil
+	case *uint:
+		n, err := util.ParseUint(b, 10, 64)
+		if err != nil {
+			return err
+		}
+		*v = uint(n)
+		return nil
+	case *uint8:
+		n, err := util.ParseUint(b, 10, 8)
+		if err != nil {
+			return err
+		}
+		*v = uint8(n)
+		return nil
+	case *uint16:
+		n, err := util.ParseUint(b, 10, 16)
+		if err != nil {
+			return err
+		}
+		*v = uint16(n)
+		return nil
+	case *uint32:
+		n, err := util.ParseUint(b, 10, 32)
+		if err != nil {
+			return err
+		}
+		*v = uint32(n)
+		return nil
+	case *uint64:
+		n, err := util.ParseUint(b, 10, 64)
+		if err != nil {
+			return err
+		}
+		*v = n
+		return nil
+	case *float32:
+		n, err := util.ParseFloat(b, 32)
+		if err != nil {
+			return err
+		}
+		*v = float32(n)
+		return err
+	case *float64:
+		var err error
+		*v, err = util.ParseFloat(b, 64)
+		return err
+	case *bool:
+		*v = len(b) == 1 && b[0] == '1'
+		return nil
+	case *time.Time:
+		var err error
+		*v, err = time.Parse(time.RFC3339Nano, util.BytesToString(b))
+		return err
+	case encoding.BinaryUnmarshaler:
+		return v.UnmarshalBinary(b)
+	default:
+		return fmt.Errorf(
+			"redis: can't unmarshal %T (consider implementing BinaryUnmarshaler)", v)
+	}
+}
+
+func ScanSlice(data []string, slice interface{}) error {
+	v := reflect.ValueOf(slice)
+	if !v.IsValid() {
+		return fmt.Errorf("redis: ScanSlice(nil)")
+	}
+	if v.Kind() != reflect.Ptr {
+		return fmt.Errorf("redis: ScanSlice(non-pointer %T)", slice)
+	}
+	v = v.Elem()
+	if v.Kind() != reflect.Slice {
+		return fmt.Errorf("redis: ScanSlice(non-slice %T)", slice)
+	}
+
+	next := makeSliceNextElemFunc(v)
+	for i, s := range data {
+		elem := next()
+		if err := Scan([]byte(s), elem.Addr().Interface()); err != nil {
+			err = fmt.Errorf("redis: ScanSlice index=%d value=%q failed: %s", i, s, err)
+			return err
+		}
+	}
+
+	return nil
+}
+
+func makeSliceNextElemFunc(v reflect.Value) func() reflect.Value {
+	elemType := v.Type().Elem()
+
+	if elemType.Kind() == reflect.Ptr {
+		elemType = elemType.Elem()
+		return func() reflect.Value {
+			if v.Len() < v.Cap() {
+				v.Set(v.Slice(0, v.Len()+1))
+				elem := v.Index(v.Len() - 1)
+				if elem.IsNil() {
+					elem.Set(reflect.New(elemType))
+				}
+				return elem.Elem()
+			}
+
+			elem := reflect.New(elemType)
+			v.Set(reflect.Append(v, elem))
+			return elem.Elem()
+		}
+	}
+
+	zero := reflect.Zero(elemType)
+	return func() reflect.Value {
+		if v.Len() < v.Cap() {
+			v.Set(v.Slice(0, v.Len()+1))
+			return v.Index(v.Len() - 1)
+		}
+
+		v.Set(reflect.Append(v, zero))
+		return v.Index(v.Len() - 1)
+	}
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/proto/writer.go b/vendor/github.com/go-redis/redis/v8/internal/proto/writer.go
new file mode 100644
index 0000000..81b09b8
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/proto/writer.go
@@ -0,0 +1,153 @@
+package proto
+
+import (
+	"encoding"
+	"fmt"
+	"io"
+	"strconv"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal/util"
+)
+
+type writer interface {
+	io.Writer
+	io.ByteWriter
+	// io.StringWriter
+	WriteString(s string) (n int, err error)
+}
+
+type Writer struct {
+	writer
+
+	lenBuf []byte
+	numBuf []byte
+}
+
+func NewWriter(wr writer) *Writer {
+	return &Writer{
+		writer: wr,
+
+		lenBuf: make([]byte, 64),
+		numBuf: make([]byte, 64),
+	}
+}
+
+func (w *Writer) WriteArgs(args []interface{}) error {
+	if err := w.WriteByte(ArrayReply); err != nil {
+		return err
+	}
+
+	if err := w.writeLen(len(args)); err != nil {
+		return err
+	}
+
+	for _, arg := range args {
+		if err := w.WriteArg(arg); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (w *Writer) writeLen(n int) error {
+	w.lenBuf = strconv.AppendUint(w.lenBuf[:0], uint64(n), 10)
+	w.lenBuf = append(w.lenBuf, '\r', '\n')
+	_, err := w.Write(w.lenBuf)
+	return err
+}
+
+func (w *Writer) WriteArg(v interface{}) error {
+	switch v := v.(type) {
+	case nil:
+		return w.string("")
+	case string:
+		return w.string(v)
+	case []byte:
+		return w.bytes(v)
+	case int:
+		return w.int(int64(v))
+	case int8:
+		return w.int(int64(v))
+	case int16:
+		return w.int(int64(v))
+	case int32:
+		return w.int(int64(v))
+	case int64:
+		return w.int(v)
+	case uint:
+		return w.uint(uint64(v))
+	case uint8:
+		return w.uint(uint64(v))
+	case uint16:
+		return w.uint(uint64(v))
+	case uint32:
+		return w.uint(uint64(v))
+	case uint64:
+		return w.uint(v)
+	case float32:
+		return w.float(float64(v))
+	case float64:
+		return w.float(v)
+	case bool:
+		if v {
+			return w.int(1)
+		}
+		return w.int(0)
+	case time.Time:
+		w.numBuf = v.AppendFormat(w.numBuf[:0], time.RFC3339Nano)
+		return w.bytes(w.numBuf)
+	case encoding.BinaryMarshaler:
+		b, err := v.MarshalBinary()
+		if err != nil {
+			return err
+		}
+		return w.bytes(b)
+	default:
+		return fmt.Errorf(
+			"redis: can't marshal %T (implement encoding.BinaryMarshaler)", v)
+	}
+}
+
+func (w *Writer) bytes(b []byte) error {
+	if err := w.WriteByte(StringReply); err != nil {
+		return err
+	}
+
+	if err := w.writeLen(len(b)); err != nil {
+		return err
+	}
+
+	if _, err := w.Write(b); err != nil {
+		return err
+	}
+
+	return w.crlf()
+}
+
+func (w *Writer) string(s string) error {
+	return w.bytes(util.StringToBytes(s))
+}
+
+func (w *Writer) uint(n uint64) error {
+	w.numBuf = strconv.AppendUint(w.numBuf[:0], n, 10)
+	return w.bytes(w.numBuf)
+}
+
+func (w *Writer) int(n int64) error {
+	w.numBuf = strconv.AppendInt(w.numBuf[:0], n, 10)
+	return w.bytes(w.numBuf)
+}
+
+func (w *Writer) float(f float64) error {
+	w.numBuf = strconv.AppendFloat(w.numBuf[:0], f, 'f', -1, 64)
+	return w.bytes(w.numBuf)
+}
+
+func (w *Writer) crlf() error {
+	if err := w.WriteByte('\r'); err != nil {
+		return err
+	}
+	return w.WriteByte('\n')
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/rand/rand.go b/vendor/github.com/go-redis/redis/v8/internal/rand/rand.go
new file mode 100644
index 0000000..40676f3
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/rand/rand.go
@@ -0,0 +1,45 @@
+package rand
+
+import (
+	"math/rand"
+	"sync"
+)
+
+// Int returns a non-negative pseudo-random int.
+func Int() int { return pseudo.Int() }
+
+// Intn returns, as an int, a non-negative pseudo-random number in [0,n).
+// It panics if n <= 0.
+func Intn(n int) int { return pseudo.Intn(n) }
+
+// Int63n returns, as an int64, a non-negative pseudo-random number in [0,n).
+// It panics if n <= 0.
+func Int63n(n int64) int64 { return pseudo.Int63n(n) }
+
+// Perm returns, as a slice of n ints, a pseudo-random permutation of the integers [0,n).
+func Perm(n int) []int { return pseudo.Perm(n) }
+
+// Seed uses the provided seed value to initialize the default Source to a
+// deterministic state. If Seed is not called, the generator behaves as if
+// seeded by Seed(1).
+func Seed(n int64) { pseudo.Seed(n) }
+
+var pseudo = rand.New(&source{src: rand.NewSource(1)})
+
+type source struct {
+	src rand.Source
+	mu  sync.Mutex
+}
+
+func (s *source) Int63() int64 {
+	s.mu.Lock()
+	n := s.src.Int63()
+	s.mu.Unlock()
+	return n
+}
+
+func (s *source) Seed(seed int64) {
+	s.mu.Lock()
+	s.src.Seed(seed)
+	s.mu.Unlock()
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/safe.go b/vendor/github.com/go-redis/redis/v8/internal/safe.go
new file mode 100644
index 0000000..862ff0e
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/safe.go
@@ -0,0 +1,11 @@
+// +build appengine
+
+package internal
+
+func String(b []byte) string {
+	return string(b)
+}
+
+func Bytes(s string) []byte {
+	return []byte(s)
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/unsafe.go b/vendor/github.com/go-redis/redis/v8/internal/unsafe.go
new file mode 100644
index 0000000..4bc7970
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/unsafe.go
@@ -0,0 +1,20 @@
+// +build !appengine
+
+package internal
+
+import "unsafe"
+
+// String converts byte slice to string.
+func String(b []byte) string {
+	return *(*string)(unsafe.Pointer(&b))
+}
+
+// Bytes converts string to byte slice.
+func Bytes(s string) []byte {
+	return *(*[]byte)(unsafe.Pointer(
+		&struct {
+			string
+			Cap int
+		}{s, len(s)},
+	))
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/util.go b/vendor/github.com/go-redis/redis/v8/internal/util.go
new file mode 100644
index 0000000..894382b
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/util.go
@@ -0,0 +1,81 @@
+package internal
+
+import (
+	"context"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal/proto"
+	"github.com/go-redis/redis/v8/internal/util"
+	"go.opentelemetry.io/otel/api/global"
+	"go.opentelemetry.io/otel/api/trace"
+)
+
+func Sleep(ctx context.Context, dur time.Duration) error {
+	return WithSpan(ctx, "time.Sleep", func(ctx context.Context, span trace.Span) error {
+		t := time.NewTimer(dur)
+		defer t.Stop()
+
+		select {
+		case <-t.C:
+			return nil
+		case <-ctx.Done():
+			return ctx.Err()
+		}
+	})
+}
+
+func ToLower(s string) string {
+	if isLower(s) {
+		return s
+	}
+
+	b := make([]byte, len(s))
+	for i := range b {
+		c := s[i]
+		if c >= 'A' && c <= 'Z' {
+			c += 'a' - 'A'
+		}
+		b[i] = c
+	}
+	return util.BytesToString(b)
+}
+
+func isLower(s string) bool {
+	for i := 0; i < len(s); i++ {
+		c := s[i]
+		if c >= 'A' && c <= 'Z' {
+			return false
+		}
+	}
+	return true
+}
+
+func Unwrap(err error) error {
+	u, ok := err.(interface {
+		Unwrap() error
+	})
+	if !ok {
+		return nil
+	}
+	return u.Unwrap()
+}
+
+//------------------------------------------------------------------------------
+
+func WithSpan(ctx context.Context, name string, fn func(context.Context, trace.Span) error) error {
+	if span := trace.SpanFromContext(ctx); !span.IsRecording() {
+		return fn(ctx, span)
+	}
+
+	ctx, span := global.Tracer("github.com/go-redis/redis").Start(ctx, name)
+	defer span.End()
+
+	return fn(ctx, span)
+}
+
+func RecordError(ctx context.Context, err error) error {
+	if err != proto.Nil {
+		trace.SpanFromContext(ctx).RecordError(ctx, err)
+	}
+	return err
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/util/safe.go b/vendor/github.com/go-redis/redis/v8/internal/util/safe.go
new file mode 100644
index 0000000..1b3060e
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/util/safe.go
@@ -0,0 +1,11 @@
+// +build appengine
+
+package util
+
+func BytesToString(b []byte) string {
+	return string(b)
+}
+
+func StringToBytes(s string) []byte {
+	return []byte(s)
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/util/strconv.go b/vendor/github.com/go-redis/redis/v8/internal/util/strconv.go
new file mode 100644
index 0000000..db50338
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/util/strconv.go
@@ -0,0 +1,19 @@
+package util
+
+import "strconv"
+
+func Atoi(b []byte) (int, error) {
+	return strconv.Atoi(BytesToString(b))
+}
+
+func ParseInt(b []byte, base int, bitSize int) (int64, error) {
+	return strconv.ParseInt(BytesToString(b), base, bitSize)
+}
+
+func ParseUint(b []byte, base int, bitSize int) (uint64, error) {
+	return strconv.ParseUint(BytesToString(b), base, bitSize)
+}
+
+func ParseFloat(b []byte, bitSize int) (float64, error) {
+	return strconv.ParseFloat(BytesToString(b), bitSize)
+}
diff --git a/vendor/github.com/go-redis/redis/v8/internal/util/unsafe.go b/vendor/github.com/go-redis/redis/v8/internal/util/unsafe.go
new file mode 100644
index 0000000..c9868aa
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/internal/util/unsafe.go
@@ -0,0 +1,22 @@
+// +build !appengine
+
+package util
+
+import (
+	"unsafe"
+)
+
+// BytesToString converts byte slice to string.
+func BytesToString(b []byte) string {
+	return *(*string)(unsafe.Pointer(&b))
+}
+
+// StringToBytes converts string to byte slice.
+func StringToBytes(s string) []byte {
+	return *(*[]byte)(unsafe.Pointer(
+		&struct {
+			string
+			Cap int
+		}{s, len(s)},
+	))
+}
diff --git a/vendor/github.com/go-redis/redis/v8/iterator.go b/vendor/github.com/go-redis/redis/v8/iterator.go
new file mode 100644
index 0000000..2f8bc2b
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/iterator.go
@@ -0,0 +1,77 @@
+package redis
+
+import (
+	"context"
+	"sync"
+)
+
+// ScanIterator is used to incrementally iterate over a collection of elements.
+// It's safe for concurrent use by multiple goroutines.
+type ScanIterator struct {
+	mu  sync.Mutex // protects Scanner and pos
+	cmd *ScanCmd
+	pos int
+}
+
+// Err returns the last iterator error, if any.
+func (it *ScanIterator) Err() error {
+	it.mu.Lock()
+	err := it.cmd.Err()
+	it.mu.Unlock()
+	return err
+}
+
+// Next advances the cursor and returns true if more values can be read.
+func (it *ScanIterator) Next(ctx context.Context) bool {
+	it.mu.Lock()
+	defer it.mu.Unlock()
+
+	// Instantly return on errors.
+	if it.cmd.Err() != nil {
+		return false
+	}
+
+	// Advance cursor, check if we are still within range.
+	if it.pos < len(it.cmd.page) {
+		it.pos++
+		return true
+	}
+
+	for {
+		// Return if there is no more data to fetch.
+		if it.cmd.cursor == 0 {
+			return false
+		}
+
+		// Fetch next page.
+		switch it.cmd.args[0] {
+		case "scan", "qscan":
+			it.cmd.args[1] = it.cmd.cursor
+		default:
+			it.cmd.args[2] = it.cmd.cursor
+		}
+
+		err := it.cmd.process(ctx, it.cmd)
+		if err != nil {
+			return false
+		}
+
+		it.pos = 1
+
+		// Redis can occasionally return empty page.
+		if len(it.cmd.page) > 0 {
+			return true
+		}
+	}
+}
+
+// Val returns the key/field at the current cursor position.
+func (it *ScanIterator) Val() string {
+	var v string
+	it.mu.Lock()
+	if it.cmd.Err() == nil && it.pos > 0 && it.pos <= len(it.cmd.page) {
+		v = it.cmd.page[it.pos-1]
+	}
+	it.mu.Unlock()
+	return v
+}
diff --git a/vendor/github.com/go-redis/redis/v8/options.go b/vendor/github.com/go-redis/redis/v8/options.go
new file mode 100644
index 0000000..f2c16c5
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/options.go
@@ -0,0 +1,317 @@
+package redis
+
+import (
+	"context"
+	"crypto/tls"
+	"errors"
+	"fmt"
+	"net"
+	"net/url"
+	"runtime"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"go.opentelemetry.io/otel/api/trace"
+	"go.opentelemetry.io/otel/label"
+)
+
+// Limiter is the interface of a rate limiter or a circuit breaker.
+type Limiter interface {
+	// Allow returns nil if operation is allowed or an error otherwise.
+	// If operation is allowed client must ReportResult of the operation
+	// whether it is a success or a failure.
+	Allow() error
+	// ReportResult reports the result of the previously allowed operation.
+	// nil indicates a success, non-nil error usually indicates a failure.
+	ReportResult(result error)
+}
+
+// Options keeps the settings to setup redis connection.
+type Options struct {
+	// The network type, either tcp or unix.
+	// Default is tcp.
+	Network string
+	// host:port address.
+	Addr string
+
+	// Dialer creates new network connection and has priority over
+	// Network and Addr options.
+	Dialer func(ctx context.Context, network, addr string) (net.Conn, error)
+
+	// Hook that is called when new connection is established.
+	OnConnect func(ctx context.Context, cn *Conn) error
+
+	// Use the specified Username to authenticate the current connection
+	// with one of the connections defined in the ACL list when connecting
+	// to a Redis 6.0 instance, or greater, that is using the Redis ACL system.
+	Username string
+	// Optional password. Must match the password specified in the
+	// requirepass server configuration option (if connecting to a Redis 5.0 instance, or lower),
+	// or the User Password when connecting to a Redis 6.0 instance, or greater,
+	// that is using the Redis ACL system.
+	Password string
+
+	// Database to be selected after connecting to the server.
+	DB int
+
+	// Maximum number of retries before giving up.
+	// Default is 3 retries.
+	MaxRetries int
+	// Minimum backoff between each retry.
+	// Default is 8 milliseconds; -1 disables backoff.
+	MinRetryBackoff time.Duration
+	// Maximum backoff between each retry.
+	// Default is 512 milliseconds; -1 disables backoff.
+	MaxRetryBackoff time.Duration
+
+	// Dial timeout for establishing new connections.
+	// Default is 5 seconds.
+	DialTimeout time.Duration
+	// Timeout for socket reads. If reached, commands will fail
+	// with a timeout instead of blocking. Use value -1 for no timeout and 0 for default.
+	// Default is 3 seconds.
+	ReadTimeout time.Duration
+	// Timeout for socket writes. If reached, commands will fail
+	// with a timeout instead of blocking.
+	// Default is ReadTimeout.
+	WriteTimeout time.Duration
+
+	// Maximum number of socket connections.
+	// Default is 10 connections per every CPU as reported by runtime.NumCPU.
+	PoolSize int
+	// Minimum number of idle connections which is useful when establishing
+	// new connection is slow.
+	MinIdleConns int
+	// Connection age at which client retires (closes) the connection.
+	// Default is to not close aged connections.
+	MaxConnAge time.Duration
+	// Amount of time client waits for connection if all connections
+	// are busy before returning an error.
+	// Default is ReadTimeout + 1 second.
+	PoolTimeout time.Duration
+	// Amount of time after which client closes idle connections.
+	// Should be less than server's timeout.
+	// Default is 5 minutes. -1 disables idle timeout check.
+	IdleTimeout time.Duration
+	// Frequency of idle checks made by idle connections reaper.
+	// Default is 1 minute. -1 disables idle connections reaper,
+	// but idle connections are still discarded by the client
+	// if IdleTimeout is set.
+	IdleCheckFrequency time.Duration
+
+	// Enables read only queries on slave nodes.
+	readOnly bool
+
+	// TLS Config to use. When set TLS will be negotiated.
+	TLSConfig *tls.Config
+
+	// Limiter interface used to implemented circuit breaker or rate limiter.
+	Limiter Limiter
+}
+
+func (opt *Options) init() {
+	if opt.Addr == "" {
+		opt.Addr = "localhost:6379"
+	}
+	if opt.Network == "" {
+		if strings.HasPrefix(opt.Addr, "/") {
+			opt.Network = "unix"
+		} else {
+			opt.Network = "tcp"
+		}
+	}
+	if opt.DialTimeout == 0 {
+		opt.DialTimeout = 5 * time.Second
+	}
+	if opt.Dialer == nil {
+		opt.Dialer = func(ctx context.Context, network, addr string) (net.Conn, error) {
+			netDialer := &net.Dialer{
+				Timeout:   opt.DialTimeout,
+				KeepAlive: 5 * time.Minute,
+			}
+			if opt.TLSConfig == nil {
+				return netDialer.DialContext(ctx, network, addr)
+			}
+			return tls.DialWithDialer(netDialer, network, addr, opt.TLSConfig)
+		}
+	}
+	if opt.PoolSize == 0 {
+		opt.PoolSize = 10 * runtime.NumCPU()
+	}
+	switch opt.ReadTimeout {
+	case -1:
+		opt.ReadTimeout = 0
+	case 0:
+		opt.ReadTimeout = 3 * time.Second
+	}
+	switch opt.WriteTimeout {
+	case -1:
+		opt.WriteTimeout = 0
+	case 0:
+		opt.WriteTimeout = opt.ReadTimeout
+	}
+	if opt.PoolTimeout == 0 {
+		opt.PoolTimeout = opt.ReadTimeout + time.Second
+	}
+	if opt.IdleTimeout == 0 {
+		opt.IdleTimeout = 5 * time.Minute
+	}
+	if opt.IdleCheckFrequency == 0 {
+		opt.IdleCheckFrequency = time.Minute
+	}
+
+	if opt.MaxRetries == -1 {
+		opt.MaxRetries = 0
+	} else if opt.MaxRetries == 0 {
+		opt.MaxRetries = 3
+	}
+	switch opt.MinRetryBackoff {
+	case -1:
+		opt.MinRetryBackoff = 0
+	case 0:
+		opt.MinRetryBackoff = 8 * time.Millisecond
+	}
+	switch opt.MaxRetryBackoff {
+	case -1:
+		opt.MaxRetryBackoff = 0
+	case 0:
+		opt.MaxRetryBackoff = 512 * time.Millisecond
+	}
+}
+
+func (opt *Options) clone() *Options {
+	clone := *opt
+	return &clone
+}
+
+// ParseURL parses an URL into Options that can be used to connect to Redis.
+// Scheme is required.
+// There are two connection types: by tcp socket and by unix socket.
+// Tcp connection:
+// 		redis://<user>:<password>@<host>:<port>/<db_number>
+// Unix connection:
+//		unix://<user>:<password>@</path/to/redis.sock>?db=<db_number>
+func ParseURL(redisURL string) (*Options, error) {
+	u, err := url.Parse(redisURL)
+	if err != nil {
+		return nil, err
+	}
+
+	switch u.Scheme {
+	case "redis", "rediss":
+		return setupTCPConn(u)
+	case "unix":
+		return setupUnixConn(u)
+	default:
+		return nil, fmt.Errorf("redis: invalid URL scheme: %s", u.Scheme)
+	}
+}
+
+func setupTCPConn(u *url.URL) (*Options, error) {
+	o := &Options{Network: "tcp"}
+
+	o.Username, o.Password = getUserPassword(u)
+
+	if len(u.Query()) > 0 {
+		return nil, errors.New("redis: no options supported")
+	}
+
+	h, p, err := net.SplitHostPort(u.Host)
+	if err != nil {
+		h = u.Host
+	}
+	if h == "" {
+		h = "localhost"
+	}
+	if p == "" {
+		p = "6379"
+	}
+	o.Addr = net.JoinHostPort(h, p)
+
+	f := strings.FieldsFunc(u.Path, func(r rune) bool {
+		return r == '/'
+	})
+	switch len(f) {
+	case 0:
+		o.DB = 0
+	case 1:
+		if o.DB, err = strconv.Atoi(f[0]); err != nil {
+			return nil, fmt.Errorf("redis: invalid database number: %q", f[0])
+		}
+	default:
+		return nil, fmt.Errorf("redis: invalid URL path: %s", u.Path)
+	}
+
+	if u.Scheme == "rediss" {
+		o.TLSConfig = &tls.Config{ServerName: h}
+	}
+
+	return o, nil
+}
+
+func setupUnixConn(u *url.URL) (*Options, error) {
+	o := &Options{
+		Network: "unix",
+	}
+
+	if strings.TrimSpace(u.Path) == "" { // path is required with unix connection
+		return nil, errors.New("redis: empty unix socket path")
+	}
+	o.Addr = u.Path
+
+	o.Username, o.Password = getUserPassword(u)
+
+	dbStr := u.Query().Get("db")
+	if dbStr == "" {
+		return o, nil // if database is not set, connect to 0 db.
+	}
+
+	db, err := strconv.Atoi(dbStr)
+	if err != nil {
+		return nil, fmt.Errorf("redis: invalid database number: %s", err)
+	}
+	o.DB = db
+
+	return o, nil
+}
+
+func getUserPassword(u *url.URL) (string, string) {
+	var user, password string
+	if u.User != nil {
+		user = u.User.Username()
+		if p, ok := u.User.Password(); ok {
+			password = p
+		}
+	}
+	return user, password
+}
+
+func newConnPool(opt *Options) *pool.ConnPool {
+	return pool.NewConnPool(&pool.Options{
+		Dialer: func(ctx context.Context) (net.Conn, error) {
+			var conn net.Conn
+			err := internal.WithSpan(ctx, "redis.dial", func(ctx context.Context, span trace.Span) error {
+				span.SetAttributes(
+					label.String("db.connection_string", opt.Addr),
+				)
+
+				var err error
+				conn, err = opt.Dialer(ctx, opt.Network, opt.Addr)
+				if err != nil {
+					_ = internal.RecordError(ctx, err)
+				}
+				return err
+			})
+			return conn, err
+		},
+		PoolSize:           opt.PoolSize,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: opt.IdleCheckFrequency,
+	})
+}
diff --git a/vendor/github.com/go-redis/redis/v8/pipeline.go b/vendor/github.com/go-redis/redis/v8/pipeline.go
new file mode 100644
index 0000000..c6ec340
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/pipeline.go
@@ -0,0 +1,137 @@
+package redis
+
+import (
+	"context"
+	"sync"
+
+	"github.com/go-redis/redis/v8/internal/pool"
+)
+
+type pipelineExecer func(context.Context, []Cmder) error
+
+// Pipeliner is an mechanism to realise Redis Pipeline technique.
+//
+// Pipelining is a technique to extremely speed up processing by packing
+// operations to batches, send them at once to Redis and read a replies in a
+// singe step.
+// See https://redis.io/topics/pipelining
+//
+// Pay attention, that Pipeline is not a transaction, so you can get unexpected
+// results in case of big pipelines and small read/write timeouts.
+// Redis client has retransmission logic in case of timeouts, pipeline
+// can be retransmitted and commands can be executed more then once.
+// To avoid this: it is good idea to use reasonable bigger read/write timeouts
+// depends of your batch size and/or use TxPipeline.
+type Pipeliner interface {
+	StatefulCmdable
+	Do(ctx context.Context, args ...interface{}) *Cmd
+	Process(ctx context.Context, cmd Cmder) error
+	Close() error
+	Discard() error
+	Exec(ctx context.Context) ([]Cmder, error)
+}
+
+var _ Pipeliner = (*Pipeline)(nil)
+
+// Pipeline implements pipelining as described in
+// http://redis.io/topics/pipelining. It's safe for concurrent use
+// by multiple goroutines.
+type Pipeline struct {
+	cmdable
+	statefulCmdable
+
+	ctx  context.Context
+	exec pipelineExecer
+
+	mu     sync.Mutex
+	cmds   []Cmder
+	closed bool
+}
+
+func (c *Pipeline) init() {
+	c.cmdable = c.Process
+	c.statefulCmdable = c.Process
+}
+
+func (c *Pipeline) Do(ctx context.Context, args ...interface{}) *Cmd {
+	cmd := NewCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Process queues the cmd for later execution.
+func (c *Pipeline) Process(ctx context.Context, cmd Cmder) error {
+	c.mu.Lock()
+	c.cmds = append(c.cmds, cmd)
+	c.mu.Unlock()
+	return nil
+}
+
+// Close closes the pipeline, releasing any open resources.
+func (c *Pipeline) Close() error {
+	c.mu.Lock()
+	_ = c.discard()
+	c.closed = true
+	c.mu.Unlock()
+	return nil
+}
+
+// Discard resets the pipeline and discards queued commands.
+func (c *Pipeline) Discard() error {
+	c.mu.Lock()
+	err := c.discard()
+	c.mu.Unlock()
+	return err
+}
+
+func (c *Pipeline) discard() error {
+	if c.closed {
+		return pool.ErrClosed
+	}
+	c.cmds = c.cmds[:0]
+	return nil
+}
+
+// Exec executes all previously queued commands using one
+// client-server roundtrip.
+//
+// Exec always returns list of commands and error of the first failed
+// command if any.
+func (c *Pipeline) Exec(ctx context.Context) ([]Cmder, error) {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.closed {
+		return nil, pool.ErrClosed
+	}
+
+	if len(c.cmds) == 0 {
+		return nil, nil
+	}
+
+	cmds := c.cmds
+	c.cmds = nil
+
+	return cmds, c.exec(ctx, cmds)
+}
+
+func (c *Pipeline) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	if err := fn(c); err != nil {
+		return nil, err
+	}
+	cmds, err := c.Exec(ctx)
+	_ = c.Close()
+	return cmds, err
+}
+
+func (c *Pipeline) Pipeline() Pipeliner {
+	return c
+}
+
+func (c *Pipeline) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipelined(ctx, fn)
+}
+
+func (c *Pipeline) TxPipeline() Pipeliner {
+	return c
+}
diff --git a/vendor/github.com/go-redis/redis/v8/pubsub.go b/vendor/github.com/go-redis/redis/v8/pubsub.go
new file mode 100644
index 0000000..c56270b
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/pubsub.go
@@ -0,0 +1,629 @@
+package redis
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/proto"
+)
+
+const (
+	pingTimeout     = time.Second
+	chanSendTimeout = time.Minute
+)
+
+var errPingTimeout = errors.New("redis: ping timeout")
+
+// PubSub implements Pub/Sub commands as described in
+// http://redis.io/topics/pubsub. Message receiving is NOT safe
+// for concurrent use by multiple goroutines.
+//
+// PubSub automatically reconnects to Redis Server and resubscribes
+// to the channels in case of network errors.
+type PubSub struct {
+	opt *Options
+
+	newConn   func(ctx context.Context, channels []string) (*pool.Conn, error)
+	closeConn func(*pool.Conn) error
+
+	mu       sync.Mutex
+	cn       *pool.Conn
+	channels map[string]struct{}
+	patterns map[string]struct{}
+
+	closed bool
+	exit   chan struct{}
+
+	cmd *Cmd
+
+	chOnce sync.Once
+	msgCh  chan *Message
+	allCh  chan interface{}
+	ping   chan struct{}
+}
+
+func (c *PubSub) String() string {
+	channels := mapKeys(c.channels)
+	channels = append(channels, mapKeys(c.patterns)...)
+	return fmt.Sprintf("PubSub(%s)", strings.Join(channels, ", "))
+}
+
+func (c *PubSub) init() {
+	c.exit = make(chan struct{})
+}
+
+func (c *PubSub) connWithLock(ctx context.Context) (*pool.Conn, error) {
+	c.mu.Lock()
+	cn, err := c.conn(ctx, nil)
+	c.mu.Unlock()
+	return cn, err
+}
+
+func (c *PubSub) conn(ctx context.Context, newChannels []string) (*pool.Conn, error) {
+	if c.closed {
+		return nil, pool.ErrClosed
+	}
+	if c.cn != nil {
+		return c.cn, nil
+	}
+
+	channels := mapKeys(c.channels)
+	channels = append(channels, newChannels...)
+
+	cn, err := c.newConn(ctx, channels)
+	if err != nil {
+		return nil, err
+	}
+
+	if err := c.resubscribe(ctx, cn); err != nil {
+		_ = c.closeConn(cn)
+		return nil, err
+	}
+
+	c.cn = cn
+	return cn, nil
+}
+
+func (c *PubSub) writeCmd(ctx context.Context, cn *pool.Conn, cmd Cmder) error {
+	return cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+		return writeCmd(wr, cmd)
+	})
+}
+
+func (c *PubSub) resubscribe(ctx context.Context, cn *pool.Conn) error {
+	var firstErr error
+
+	if len(c.channels) > 0 {
+		firstErr = c._subscribe(ctx, cn, "subscribe", mapKeys(c.channels))
+	}
+
+	if len(c.patterns) > 0 {
+		err := c._subscribe(ctx, cn, "psubscribe", mapKeys(c.patterns))
+		if err != nil && firstErr == nil {
+			firstErr = err
+		}
+	}
+
+	return firstErr
+}
+
+func mapKeys(m map[string]struct{}) []string {
+	s := make([]string, len(m))
+	i := 0
+	for k := range m {
+		s[i] = k
+		i++
+	}
+	return s
+}
+
+func (c *PubSub) _subscribe(
+	ctx context.Context, cn *pool.Conn, redisCmd string, channels []string,
+) error {
+	args := make([]interface{}, 0, 1+len(channels))
+	args = append(args, redisCmd)
+	for _, channel := range channels {
+		args = append(args, channel)
+	}
+	cmd := NewSliceCmd(ctx, args...)
+	return c.writeCmd(ctx, cn, cmd)
+}
+
+func (c *PubSub) releaseConnWithLock(
+	ctx context.Context,
+	cn *pool.Conn,
+	err error,
+	allowTimeout bool,
+) {
+	c.mu.Lock()
+	c.releaseConn(ctx, cn, err, allowTimeout)
+	c.mu.Unlock()
+}
+
+func (c *PubSub) releaseConn(ctx context.Context, cn *pool.Conn, err error, allowTimeout bool) {
+	if c.cn != cn {
+		return
+	}
+	if isBadConn(err, allowTimeout) {
+		c.reconnect(ctx, err)
+	}
+}
+
+func (c *PubSub) reconnect(ctx context.Context, reason error) {
+	_ = c.closeTheCn(reason)
+	_, _ = c.conn(ctx, nil)
+}
+
+func (c *PubSub) closeTheCn(reason error) error {
+	if c.cn == nil {
+		return nil
+	}
+	if !c.closed {
+		internal.Logger.Printf(c.getContext(), "redis: discarding bad PubSub connection: %s", reason)
+	}
+	err := c.closeConn(c.cn)
+	c.cn = nil
+	return err
+}
+
+func (c *PubSub) Close() error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.closed {
+		return pool.ErrClosed
+	}
+	c.closed = true
+	close(c.exit)
+
+	return c.closeTheCn(pool.ErrClosed)
+}
+
+// Subscribe the client to the specified channels. It returns
+// empty subscription if there are no channels.
+func (c *PubSub) Subscribe(ctx context.Context, channels ...string) error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	err := c.subscribe(ctx, "subscribe", channels...)
+	if c.channels == nil {
+		c.channels = make(map[string]struct{})
+	}
+	for _, s := range channels {
+		c.channels[s] = struct{}{}
+	}
+	return err
+}
+
+// PSubscribe the client to the given patterns. It returns
+// empty subscription if there are no patterns.
+func (c *PubSub) PSubscribe(ctx context.Context, patterns ...string) error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	err := c.subscribe(ctx, "psubscribe", patterns...)
+	if c.patterns == nil {
+		c.patterns = make(map[string]struct{})
+	}
+	for _, s := range patterns {
+		c.patterns[s] = struct{}{}
+	}
+	return err
+}
+
+// Unsubscribe the client from the given channels, or from all of
+// them if none is given.
+func (c *PubSub) Unsubscribe(ctx context.Context, channels ...string) error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	for _, channel := range channels {
+		delete(c.channels, channel)
+	}
+	err := c.subscribe(ctx, "unsubscribe", channels...)
+	return err
+}
+
+// PUnsubscribe the client from the given patterns, or from all of
+// them if none is given.
+func (c *PubSub) PUnsubscribe(ctx context.Context, patterns ...string) error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	for _, pattern := range patterns {
+		delete(c.patterns, pattern)
+	}
+	err := c.subscribe(ctx, "punsubscribe", patterns...)
+	return err
+}
+
+func (c *PubSub) subscribe(ctx context.Context, redisCmd string, channels ...string) error {
+	cn, err := c.conn(ctx, channels)
+	if err != nil {
+		return err
+	}
+
+	err = c._subscribe(ctx, cn, redisCmd, channels)
+	c.releaseConn(ctx, cn, err, false)
+	return err
+}
+
+func (c *PubSub) Ping(ctx context.Context, payload ...string) error {
+	args := []interface{}{"ping"}
+	if len(payload) == 1 {
+		args = append(args, payload[0])
+	}
+	cmd := NewCmd(ctx, args...)
+
+	cn, err := c.connWithLock(ctx)
+	if err != nil {
+		return err
+	}
+
+	err = c.writeCmd(ctx, cn, cmd)
+	c.releaseConnWithLock(ctx, cn, err, false)
+	return err
+}
+
+// Subscription received after a successful subscription to channel.
+type Subscription struct {
+	// Can be "subscribe", "unsubscribe", "psubscribe" or "punsubscribe".
+	Kind string
+	// Channel name we have subscribed to.
+	Channel string
+	// Number of channels we are currently subscribed to.
+	Count int
+}
+
+func (m *Subscription) String() string {
+	return fmt.Sprintf("%s: %s", m.Kind, m.Channel)
+}
+
+// Message received as result of a PUBLISH command issued by another client.
+type Message struct {
+	Channel      string
+	Pattern      string
+	Payload      string
+	PayloadSlice []string
+}
+
+func (m *Message) String() string {
+	return fmt.Sprintf("Message<%s: %s>", m.Channel, m.Payload)
+}
+
+// Pong received as result of a PING command issued by another client.
+type Pong struct {
+	Payload string
+}
+
+func (p *Pong) String() string {
+	if p.Payload != "" {
+		return fmt.Sprintf("Pong<%s>", p.Payload)
+	}
+	return "Pong"
+}
+
+func (c *PubSub) newMessage(reply interface{}) (interface{}, error) {
+	switch reply := reply.(type) {
+	case string:
+		return &Pong{
+			Payload: reply,
+		}, nil
+	case []interface{}:
+		switch kind := reply[0].(string); kind {
+		case "subscribe", "unsubscribe", "psubscribe", "punsubscribe":
+			// Can be nil in case of "unsubscribe".
+			channel, _ := reply[1].(string)
+			return &Subscription{
+				Kind:    kind,
+				Channel: channel,
+				Count:   int(reply[2].(int64)),
+			}, nil
+		case "message":
+			switch payload := reply[2].(type) {
+			case string:
+				return &Message{
+					Channel: reply[1].(string),
+					Payload: payload,
+				}, nil
+			case []interface{}:
+				ss := make([]string, len(payload))
+				for i, s := range payload {
+					ss[i] = s.(string)
+				}
+				return &Message{
+					Channel:      reply[1].(string),
+					PayloadSlice: ss,
+				}, nil
+			default:
+				return nil, fmt.Errorf("redis: unsupported pubsub message payload: %T", payload)
+			}
+		case "pmessage":
+			return &Message{
+				Pattern: reply[1].(string),
+				Channel: reply[2].(string),
+				Payload: reply[3].(string),
+			}, nil
+		case "pong":
+			return &Pong{
+				Payload: reply[1].(string),
+			}, nil
+		default:
+			return nil, fmt.Errorf("redis: unsupported pubsub message: %q", kind)
+		}
+	default:
+		return nil, fmt.Errorf("redis: unsupported pubsub message: %#v", reply)
+	}
+}
+
+// ReceiveTimeout acts like Receive but returns an error if message
+// is not received in time. This is low-level API and in most cases
+// Channel should be used instead.
+func (c *PubSub) ReceiveTimeout(ctx context.Context, timeout time.Duration) (interface{}, error) {
+	if c.cmd == nil {
+		c.cmd = NewCmd(ctx)
+	}
+
+	cn, err := c.connWithLock(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	err = cn.WithReader(ctx, timeout, func(rd *proto.Reader) error {
+		return c.cmd.readReply(rd)
+	})
+
+	c.releaseConnWithLock(ctx, cn, err, timeout > 0)
+	if err != nil {
+		return nil, err
+	}
+
+	return c.newMessage(c.cmd.Val())
+}
+
+// Receive returns a message as a Subscription, Message, Pong or error.
+// See PubSub example for details. This is low-level API and in most cases
+// Channel should be used instead.
+func (c *PubSub) Receive(ctx context.Context) (interface{}, error) {
+	return c.ReceiveTimeout(ctx, 0)
+}
+
+// ReceiveMessage returns a Message or error ignoring Subscription and Pong
+// messages. This is low-level API and in most cases Channel should be used
+// instead.
+func (c *PubSub) ReceiveMessage(ctx context.Context) (*Message, error) {
+	for {
+		msg, err := c.Receive(ctx)
+		if err != nil {
+			return nil, err
+		}
+
+		switch msg := msg.(type) {
+		case *Subscription:
+			// Ignore.
+		case *Pong:
+			// Ignore.
+		case *Message:
+			return msg, nil
+		default:
+			err := fmt.Errorf("redis: unknown message: %T", msg)
+			return nil, err
+		}
+	}
+}
+
+// Channel returns a Go channel for concurrently receiving messages.
+// The channel is closed together with the PubSub. If the Go channel
+// is blocked full for 30 seconds the message is dropped.
+// Receive* APIs can not be used after channel is created.
+//
+// go-redis periodically sends ping messages to test connection health
+// and re-subscribes if ping can not not received for 30 seconds.
+func (c *PubSub) Channel() <-chan *Message {
+	return c.ChannelSize(100)
+}
+
+// ChannelSize is like Channel, but creates a Go channel
+// with specified buffer size.
+func (c *PubSub) ChannelSize(size int) <-chan *Message {
+	c.chOnce.Do(func() {
+		c.initPing()
+		c.initMsgChan(size)
+	})
+	if c.msgCh == nil {
+		err := fmt.Errorf("redis: Channel can't be called after ChannelWithSubscriptions")
+		panic(err)
+	}
+	if cap(c.msgCh) != size {
+		err := fmt.Errorf("redis: PubSub.Channel size can not be changed once created")
+		panic(err)
+	}
+	return c.msgCh
+}
+
+// ChannelWithSubscriptions is like Channel, but message type can be either
+// *Subscription or *Message. Subscription messages can be used to detect
+// reconnections.
+//
+// ChannelWithSubscriptions can not be used together with Channel or ChannelSize.
+func (c *PubSub) ChannelWithSubscriptions(ctx context.Context, size int) <-chan interface{} {
+	c.chOnce.Do(func() {
+		c.initPing()
+		c.initAllChan(size)
+	})
+	if c.allCh == nil {
+		err := fmt.Errorf("redis: ChannelWithSubscriptions can't be called after Channel")
+		panic(err)
+	}
+	if cap(c.allCh) != size {
+		err := fmt.Errorf("redis: PubSub.Channel size can not be changed once created")
+		panic(err)
+	}
+	return c.allCh
+}
+
+func (c *PubSub) getContext() context.Context {
+	if c.cmd != nil {
+		return c.cmd.ctx
+	}
+	return context.Background()
+}
+
+func (c *PubSub) initPing() {
+	ctx := context.TODO()
+	c.ping = make(chan struct{}, 1)
+	go func() {
+		timer := time.NewTimer(time.Minute)
+		timer.Stop()
+
+		healthy := true
+		for {
+			timer.Reset(pingTimeout)
+			select {
+			case <-c.ping:
+				healthy = true
+				if !timer.Stop() {
+					<-timer.C
+				}
+			case <-timer.C:
+				pingErr := c.Ping(ctx)
+				if healthy {
+					healthy = false
+				} else {
+					if pingErr == nil {
+						pingErr = errPingTimeout
+					}
+					c.mu.Lock()
+					c.reconnect(ctx, pingErr)
+					healthy = true
+					c.mu.Unlock()
+				}
+			case <-c.exit:
+				return
+			}
+		}
+	}()
+}
+
+// initMsgChan must be in sync with initAllChan.
+func (c *PubSub) initMsgChan(size int) {
+	ctx := context.TODO()
+	c.msgCh = make(chan *Message, size)
+	go func() {
+		timer := time.NewTimer(time.Minute)
+		timer.Stop()
+
+		var errCount int
+		for {
+			msg, err := c.Receive(ctx)
+			if err != nil {
+				if err == pool.ErrClosed {
+					close(c.msgCh)
+					return
+				}
+				if errCount > 0 {
+					time.Sleep(100 * time.Millisecond)
+				}
+				errCount++
+				continue
+			}
+
+			errCount = 0
+
+			// Any message is as good as a ping.
+			select {
+			case c.ping <- struct{}{}:
+			default:
+			}
+
+			switch msg := msg.(type) {
+			case *Subscription:
+				// Ignore.
+			case *Pong:
+				// Ignore.
+			case *Message:
+				timer.Reset(chanSendTimeout)
+				select {
+				case c.msgCh <- msg:
+					if !timer.Stop() {
+						<-timer.C
+					}
+				case <-timer.C:
+					internal.Logger.Printf(
+						c.getContext(),
+						"redis: %s channel is full for %s (message is dropped)",
+						c,
+						chanSendTimeout,
+					)
+				}
+			default:
+				internal.Logger.Printf(c.getContext(), "redis: unknown message type: %T", msg)
+			}
+		}
+	}()
+}
+
+// initAllChan must be in sync with initMsgChan.
+func (c *PubSub) initAllChan(size int) {
+	ctx := context.TODO()
+	c.allCh = make(chan interface{}, size)
+	go func() {
+		timer := time.NewTimer(pingTimeout)
+		timer.Stop()
+
+		var errCount int
+		for {
+			msg, err := c.Receive(ctx)
+			if err != nil {
+				if err == pool.ErrClosed {
+					close(c.allCh)
+					return
+				}
+				if errCount > 0 {
+					time.Sleep(100 * time.Millisecond)
+				}
+				errCount++
+				continue
+			}
+
+			errCount = 0
+
+			// Any message is as good as a ping.
+			select {
+			case c.ping <- struct{}{}:
+			default:
+			}
+
+			switch msg := msg.(type) {
+			case *Subscription:
+				c.sendMessage(msg, timer)
+			case *Pong:
+				// Ignore.
+			case *Message:
+				c.sendMessage(msg, timer)
+			default:
+				internal.Logger.Printf(c.getContext(), "redis: unknown message type: %T", msg)
+			}
+		}
+	}()
+}
+
+func (c *PubSub) sendMessage(msg interface{}, timer *time.Timer) {
+	timer.Reset(pingTimeout)
+	select {
+	case c.allCh <- msg:
+		if !timer.Stop() {
+			<-timer.C
+		}
+	case <-timer.C:
+		internal.Logger.Printf(
+			c.getContext(),
+			"redis: %s channel is full for %s (message is dropped)", c, pingTimeout)
+	}
+}
diff --git a/vendor/github.com/go-redis/redis/v8/redis.go b/vendor/github.com/go-redis/redis/v8/redis.go
new file mode 100644
index 0000000..efad7f1
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/redis.go
@@ -0,0 +1,783 @@
+package redis
+
+import (
+	"context"
+	"fmt"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/proto"
+	"go.opentelemetry.io/otel/api/trace"
+	"go.opentelemetry.io/otel/label"
+)
+
+// Nil reply returned by Redis when key does not exist.
+const Nil = proto.Nil
+
+func SetLogger(logger internal.Logging) {
+	internal.Logger = logger
+}
+
+//------------------------------------------------------------------------------
+
+type Hook interface {
+	BeforeProcess(ctx context.Context, cmd Cmder) (context.Context, error)
+	AfterProcess(ctx context.Context, cmd Cmder) error
+
+	BeforeProcessPipeline(ctx context.Context, cmds []Cmder) (context.Context, error)
+	AfterProcessPipeline(ctx context.Context, cmds []Cmder) error
+}
+
+type hooks struct {
+	hooks []Hook
+}
+
+func (hs *hooks) lock() {
+	hs.hooks = hs.hooks[:len(hs.hooks):len(hs.hooks)]
+}
+
+func (hs hooks) clone() hooks {
+	clone := hs
+	clone.lock()
+	return clone
+}
+
+func (hs *hooks) AddHook(hook Hook) {
+	hs.hooks = append(hs.hooks, hook)
+}
+
+func (hs hooks) process(
+	ctx context.Context, cmd Cmder, fn func(context.Context, Cmder) error,
+) error {
+	if len(hs.hooks) == 0 {
+		err := hs.withContext(ctx, func() error {
+			return fn(ctx, cmd)
+		})
+		cmd.SetErr(err)
+		return err
+	}
+
+	var hookIndex int
+	var retErr error
+
+	for ; hookIndex < len(hs.hooks) && retErr == nil; hookIndex++ {
+		ctx, retErr = hs.hooks[hookIndex].BeforeProcess(ctx, cmd)
+		if retErr != nil {
+			cmd.SetErr(retErr)
+		}
+	}
+
+	if retErr == nil {
+		retErr = hs.withContext(ctx, func() error {
+			return fn(ctx, cmd)
+		})
+		cmd.SetErr(retErr)
+	}
+
+	for hookIndex--; hookIndex >= 0; hookIndex-- {
+		if err := hs.hooks[hookIndex].AfterProcess(ctx, cmd); err != nil {
+			retErr = err
+			cmd.SetErr(retErr)
+		}
+	}
+
+	return retErr
+}
+
+func (hs hooks) processPipeline(
+	ctx context.Context, cmds []Cmder, fn func(context.Context, []Cmder) error,
+) error {
+	if len(hs.hooks) == 0 {
+		err := hs.withContext(ctx, func() error {
+			return fn(ctx, cmds)
+		})
+		return err
+	}
+
+	var hookIndex int
+	var retErr error
+
+	for ; hookIndex < len(hs.hooks) && retErr == nil; hookIndex++ {
+		ctx, retErr = hs.hooks[hookIndex].BeforeProcessPipeline(ctx, cmds)
+		if retErr != nil {
+			setCmdsErr(cmds, retErr)
+		}
+	}
+
+	if retErr == nil {
+		retErr = hs.withContext(ctx, func() error {
+			return fn(ctx, cmds)
+		})
+	}
+
+	for hookIndex--; hookIndex >= 0; hookIndex-- {
+		if err := hs.hooks[hookIndex].AfterProcessPipeline(ctx, cmds); err != nil {
+			retErr = err
+			setCmdsErr(cmds, retErr)
+		}
+	}
+
+	return retErr
+}
+
+func (hs hooks) processTxPipeline(
+	ctx context.Context, cmds []Cmder, fn func(context.Context, []Cmder) error,
+) error {
+	cmds = wrapMultiExec(ctx, cmds)
+	return hs.processPipeline(ctx, cmds, fn)
+}
+
+func (hs hooks) withContext(ctx context.Context, fn func() error) error {
+	done := ctx.Done()
+	if done == nil {
+		return fn()
+	}
+
+	errc := make(chan error, 1)
+	go func() { errc <- fn() }()
+
+	select {
+	case <-done:
+		return ctx.Err()
+	case err := <-errc:
+		return err
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type baseClient struct {
+	opt      *Options
+	connPool pool.Pooler
+
+	onClose func() error // hook called when client is closed
+}
+
+func newBaseClient(opt *Options, connPool pool.Pooler) *baseClient {
+	return &baseClient{
+		opt:      opt,
+		connPool: connPool,
+	}
+}
+
+func (c *baseClient) clone() *baseClient {
+	clone := *c
+	return &clone
+}
+
+func (c *baseClient) withTimeout(timeout time.Duration) *baseClient {
+	opt := c.opt.clone()
+	opt.ReadTimeout = timeout
+	opt.WriteTimeout = timeout
+
+	clone := c.clone()
+	clone.opt = opt
+
+	return clone
+}
+
+func (c *baseClient) String() string {
+	return fmt.Sprintf("Redis<%s db:%d>", c.getAddr(), c.opt.DB)
+}
+
+func (c *baseClient) newConn(ctx context.Context) (*pool.Conn, error) {
+	cn, err := c.connPool.NewConn(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	err = c.initConn(ctx, cn)
+	if err != nil {
+		_ = c.connPool.CloseConn(cn)
+		return nil, err
+	}
+
+	return cn, nil
+}
+
+func (c *baseClient) getConn(ctx context.Context) (*pool.Conn, error) {
+	if c.opt.Limiter != nil {
+		err := c.opt.Limiter.Allow()
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	cn, err := c._getConn(ctx)
+	if err != nil {
+		if c.opt.Limiter != nil {
+			c.opt.Limiter.ReportResult(err)
+		}
+		return nil, err
+	}
+
+	return cn, nil
+}
+
+func (c *baseClient) _getConn(ctx context.Context) (*pool.Conn, error) {
+	cn, err := c.connPool.Get(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	if cn.Inited {
+		return cn, nil
+	}
+
+	err = internal.WithSpan(ctx, "redis.init_conn", func(ctx context.Context, span trace.Span) error {
+		return c.initConn(ctx, cn)
+	})
+	if err != nil {
+		c.connPool.Remove(ctx, cn, err)
+		if err := internal.Unwrap(err); err != nil {
+			return nil, err
+		}
+		return nil, err
+	}
+
+	return cn, nil
+}
+
+func (c *baseClient) initConn(ctx context.Context, cn *pool.Conn) error {
+	if cn.Inited {
+		return nil
+	}
+	cn.Inited = true
+
+	if c.opt.Password == "" &&
+		c.opt.DB == 0 &&
+		!c.opt.readOnly &&
+		c.opt.OnConnect == nil {
+		return nil
+	}
+
+	connPool := pool.NewSingleConnPool(c.connPool, cn)
+	conn := newConn(ctx, c.opt, connPool)
+
+	_, err := conn.Pipelined(ctx, func(pipe Pipeliner) error {
+		if c.opt.Password != "" {
+			if c.opt.Username != "" {
+				pipe.AuthACL(ctx, c.opt.Username, c.opt.Password)
+			} else {
+				pipe.Auth(ctx, c.opt.Password)
+			}
+		}
+
+		if c.opt.DB > 0 {
+			pipe.Select(ctx, c.opt.DB)
+		}
+
+		if c.opt.readOnly {
+			pipe.ReadOnly(ctx)
+		}
+
+		return nil
+	})
+	if err != nil {
+		return err
+	}
+
+	if c.opt.OnConnect != nil {
+		return c.opt.OnConnect(ctx, conn)
+	}
+	return nil
+}
+
+func (c *baseClient) releaseConn(ctx context.Context, cn *pool.Conn, err error) {
+	if c.opt.Limiter != nil {
+		c.opt.Limiter.ReportResult(err)
+	}
+
+	if isBadConn(err, false) {
+		c.connPool.Remove(ctx, cn, err)
+	} else {
+		c.connPool.Put(ctx, cn)
+	}
+}
+
+func (c *baseClient) withConn(
+	ctx context.Context, fn func(context.Context, *pool.Conn) error,
+) error {
+	return internal.WithSpan(ctx, "redis.with_conn", func(ctx context.Context, span trace.Span) error {
+		cn, err := c.getConn(ctx)
+		if err != nil {
+			return err
+		}
+
+		if span.IsRecording() {
+			if remoteAddr := cn.RemoteAddr(); remoteAddr != nil {
+				span.SetAttributes(label.String("net.peer.ip", remoteAddr.String()))
+			}
+		}
+
+		defer func() {
+			c.releaseConn(ctx, cn, err)
+		}()
+
+		err = fn(ctx, cn)
+		return err
+	})
+}
+
+func (c *baseClient) process(ctx context.Context, cmd Cmder) error {
+	var lastErr error
+	for attempt := 0; attempt <= c.opt.MaxRetries; attempt++ {
+		attempt := attempt
+
+		var retry bool
+		err := internal.WithSpan(ctx, "redis.process", func(ctx context.Context, span trace.Span) error {
+			if attempt > 0 {
+				if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+					return err
+				}
+			}
+
+			retryTimeout := true
+			err := c.withConn(ctx, func(ctx context.Context, cn *pool.Conn) error {
+				err := cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+					return writeCmd(wr, cmd)
+				})
+				if err != nil {
+					return err
+				}
+
+				err = cn.WithReader(ctx, c.cmdTimeout(cmd), cmd.readReply)
+				if err != nil {
+					retryTimeout = cmd.readTimeout() == nil
+					return err
+				}
+
+				return nil
+			})
+			if err == nil {
+				return nil
+			}
+			retry = shouldRetry(err, retryTimeout)
+			return err
+		})
+		if err == nil || !retry {
+			return err
+		}
+		lastErr = err
+	}
+	return lastErr
+}
+
+func (c *baseClient) retryBackoff(attempt int) time.Duration {
+	return internal.RetryBackoff(attempt, c.opt.MinRetryBackoff, c.opt.MaxRetryBackoff)
+}
+
+func (c *baseClient) cmdTimeout(cmd Cmder) time.Duration {
+	if timeout := cmd.readTimeout(); timeout != nil {
+		t := *timeout
+		if t == 0 {
+			return 0
+		}
+		return t + 10*time.Second
+	}
+	return c.opt.ReadTimeout
+}
+
+// Close closes the client, releasing any open resources.
+//
+// It is rare to Close a Client, as the Client is meant to be
+// long-lived and shared between many goroutines.
+func (c *baseClient) Close() error {
+	var firstErr error
+	if c.onClose != nil {
+		if err := c.onClose(); err != nil {
+			firstErr = err
+		}
+	}
+	if err := c.connPool.Close(); err != nil && firstErr == nil {
+		firstErr = err
+	}
+	return firstErr
+}
+
+func (c *baseClient) getAddr() string {
+	return c.opt.Addr
+}
+
+func (c *baseClient) processPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.generalProcessPipeline(ctx, cmds, c.pipelineProcessCmds)
+}
+
+func (c *baseClient) processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.generalProcessPipeline(ctx, cmds, c.txPipelineProcessCmds)
+}
+
+type pipelineProcessor func(context.Context, *pool.Conn, []Cmder) (bool, error)
+
+func (c *baseClient) generalProcessPipeline(
+	ctx context.Context, cmds []Cmder, p pipelineProcessor,
+) error {
+	err := c._generalProcessPipeline(ctx, cmds, p)
+	if err != nil {
+		setCmdsErr(cmds, err)
+		return err
+	}
+	return cmdsFirstErr(cmds)
+}
+
+func (c *baseClient) _generalProcessPipeline(
+	ctx context.Context, cmds []Cmder, p pipelineProcessor,
+) error {
+	var lastErr error
+	for attempt := 0; attempt <= c.opt.MaxRetries; attempt++ {
+		if attempt > 0 {
+			if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+				return err
+			}
+		}
+
+		var canRetry bool
+		lastErr = c.withConn(ctx, func(ctx context.Context, cn *pool.Conn) error {
+			var err error
+			canRetry, err = p(ctx, cn, cmds)
+			return err
+		})
+		if lastErr == nil || !canRetry || !shouldRetry(lastErr, true) {
+			return lastErr
+		}
+	}
+	return lastErr
+}
+
+func (c *baseClient) pipelineProcessCmds(
+	ctx context.Context, cn *pool.Conn, cmds []Cmder,
+) (bool, error) {
+	err := cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+		return writeCmds(wr, cmds)
+	})
+	if err != nil {
+		return true, err
+	}
+
+	err = cn.WithReader(ctx, c.opt.ReadTimeout, func(rd *proto.Reader) error {
+		return pipelineReadCmds(rd, cmds)
+	})
+	return true, err
+}
+
+func pipelineReadCmds(rd *proto.Reader, cmds []Cmder) error {
+	for _, cmd := range cmds {
+		err := cmd.readReply(rd)
+		cmd.SetErr(err)
+		if err != nil && !isRedisError(err) {
+			return err
+		}
+	}
+	return nil
+}
+
+func (c *baseClient) txPipelineProcessCmds(
+	ctx context.Context, cn *pool.Conn, cmds []Cmder,
+) (bool, error) {
+	err := cn.WithWriter(ctx, c.opt.WriteTimeout, func(wr *proto.Writer) error {
+		return writeCmds(wr, cmds)
+	})
+	if err != nil {
+		return true, err
+	}
+
+	err = cn.WithReader(ctx, c.opt.ReadTimeout, func(rd *proto.Reader) error {
+		statusCmd := cmds[0].(*StatusCmd)
+		// Trim multi and exec.
+		cmds = cmds[1 : len(cmds)-1]
+
+		err := txPipelineReadQueued(rd, statusCmd, cmds)
+		if err != nil {
+			return err
+		}
+
+		return pipelineReadCmds(rd, cmds)
+	})
+	return false, err
+}
+
+func wrapMultiExec(ctx context.Context, cmds []Cmder) []Cmder {
+	if len(cmds) == 0 {
+		panic("not reached")
+	}
+	cmdCopy := make([]Cmder, len(cmds)+2)
+	cmdCopy[0] = NewStatusCmd(ctx, "multi")
+	copy(cmdCopy[1:], cmds)
+	cmdCopy[len(cmdCopy)-1] = NewSliceCmd(ctx, "exec")
+	return cmdCopy
+}
+
+func txPipelineReadQueued(rd *proto.Reader, statusCmd *StatusCmd, cmds []Cmder) error {
+	// Parse queued replies.
+	if err := statusCmd.readReply(rd); err != nil {
+		return err
+	}
+
+	for range cmds {
+		if err := statusCmd.readReply(rd); err != nil && !isRedisError(err) {
+			return err
+		}
+	}
+
+	// Parse number of replies.
+	line, err := rd.ReadLine()
+	if err != nil {
+		if err == Nil {
+			err = TxFailedErr
+		}
+		return err
+	}
+
+	switch line[0] {
+	case proto.ErrorReply:
+		return proto.ParseErrorReply(line)
+	case proto.ArrayReply:
+		// ok
+	default:
+		err := fmt.Errorf("redis: expected '*', but got line %q", line)
+		return err
+	}
+
+	return nil
+}
+
+//------------------------------------------------------------------------------
+
+// Client is a Redis client representing a pool of zero or more
+// underlying connections. It's safe for concurrent use by multiple
+// goroutines.
+type Client struct {
+	*baseClient
+	cmdable
+	hooks
+	ctx context.Context
+}
+
+// NewClient returns a client to the Redis Server specified by Options.
+func NewClient(opt *Options) *Client {
+	opt.init()
+
+	c := Client{
+		baseClient: newBaseClient(opt, newConnPool(opt)),
+		ctx:        context.Background(),
+	}
+	c.cmdable = c.Process
+
+	return &c
+}
+
+func (c *Client) clone() *Client {
+	clone := *c
+	clone.cmdable = clone.Process
+	clone.hooks.lock()
+	return &clone
+}
+
+func (c *Client) WithTimeout(timeout time.Duration) *Client {
+	clone := c.clone()
+	clone.baseClient = c.baseClient.withTimeout(timeout)
+	return clone
+}
+
+func (c *Client) Context() context.Context {
+	return c.ctx
+}
+
+func (c *Client) WithContext(ctx context.Context) *Client {
+	if ctx == nil {
+		panic("nil context")
+	}
+	clone := c.clone()
+	clone.ctx = ctx
+	return clone
+}
+
+func (c *Client) Conn(ctx context.Context) *Conn {
+	return newConn(ctx, c.opt, pool.NewStickyConnPool(c.connPool))
+}
+
+// Do creates a Cmd from the args and processes the cmd.
+func (c *Client) Do(ctx context.Context, args ...interface{}) *Cmd {
+	cmd := NewCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+func (c *Client) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.baseClient.process)
+}
+
+func (c *Client) processPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, c.baseClient.processPipeline)
+}
+
+func (c *Client) processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processTxPipeline(ctx, cmds, c.baseClient.processTxPipeline)
+}
+
+// Options returns read-only Options that were used to create the client.
+func (c *Client) Options() *Options {
+	return c.opt
+}
+
+type PoolStats pool.Stats
+
+// PoolStats returns connection pool stats.
+func (c *Client) PoolStats() *PoolStats {
+	stats := c.connPool.Stats()
+	return (*PoolStats)(stats)
+}
+
+func (c *Client) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipeline().Pipelined(ctx, fn)
+}
+
+func (c *Client) Pipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *Client) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.TxPipeline().Pipelined(ctx, fn)
+}
+
+// TxPipeline acts like Pipeline, but wraps queued commands with MULTI/EXEC.
+func (c *Client) TxPipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processTxPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *Client) pubSub() *PubSub {
+	pubsub := &PubSub{
+		opt: c.opt,
+
+		newConn: func(ctx context.Context, channels []string) (*pool.Conn, error) {
+			return c.newConn(ctx)
+		},
+		closeConn: c.connPool.CloseConn,
+	}
+	pubsub.init()
+	return pubsub
+}
+
+// Subscribe subscribes the client to the specified channels.
+// Channels can be omitted to create empty subscription.
+// Note that this method does not wait on a response from Redis, so the
+// subscription may not be active immediately. To force the connection to wait,
+// you may call the Receive() method on the returned *PubSub like so:
+//
+//    sub := client.Subscribe(queryResp)
+//    iface, err := sub.Receive()
+//    if err != nil {
+//        // handle error
+//    }
+//
+//    // Should be *Subscription, but others are possible if other actions have been
+//    // taken on sub since it was created.
+//    switch iface.(type) {
+//    case *Subscription:
+//        // subscribe succeeded
+//    case *Message:
+//        // received first message
+//    case *Pong:
+//        // pong received
+//    default:
+//        // handle error
+//    }
+//
+//    ch := sub.Channel()
+func (c *Client) Subscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.Subscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+// PSubscribe subscribes the client to the given patterns.
+// Patterns can be omitted to create empty subscription.
+func (c *Client) PSubscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.PSubscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+//------------------------------------------------------------------------------
+
+type conn struct {
+	baseClient
+	cmdable
+	statefulCmdable
+	hooks // TODO: inherit hooks
+}
+
+// Conn is like Client, but its pool contains single connection.
+type Conn struct {
+	*conn
+	ctx context.Context
+}
+
+func newConn(ctx context.Context, opt *Options, connPool pool.Pooler) *Conn {
+	c := Conn{
+		conn: &conn{
+			baseClient: baseClient{
+				opt:      opt,
+				connPool: connPool,
+			},
+		},
+		ctx: ctx,
+	}
+	c.cmdable = c.Process
+	c.statefulCmdable = c.Process
+	return &c
+}
+
+func (c *Conn) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.baseClient.process)
+}
+
+func (c *Conn) processPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, c.baseClient.processPipeline)
+}
+
+func (c *Conn) processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processTxPipeline(ctx, cmds, c.baseClient.processTxPipeline)
+}
+
+func (c *Conn) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipeline().Pipelined(ctx, fn)
+}
+
+func (c *Conn) Pipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *Conn) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.TxPipeline().Pipelined(ctx, fn)
+}
+
+// TxPipeline acts like Pipeline, but wraps queued commands with MULTI/EXEC.
+func (c *Conn) TxPipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processTxPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
diff --git a/vendor/github.com/go-redis/redis/v8/renovate.json b/vendor/github.com/go-redis/redis/v8/renovate.json
new file mode 100644
index 0000000..f45d8f1
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/renovate.json
@@ -0,0 +1,5 @@
+{
+  "extends": [
+    "config:base"
+  ]
+}
diff --git a/vendor/github.com/go-redis/redis/v8/result.go b/vendor/github.com/go-redis/redis/v8/result.go
new file mode 100644
index 0000000..24cfd49
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/result.go
@@ -0,0 +1,180 @@
+package redis
+
+import "time"
+
+// NewCmdResult returns a Cmd initialised with val and err for testing.
+func NewCmdResult(val interface{}, err error) *Cmd {
+	var cmd Cmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewSliceResult returns a SliceCmd initialised with val and err for testing.
+func NewSliceResult(val []interface{}, err error) *SliceCmd {
+	var cmd SliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewStatusResult returns a StatusCmd initialised with val and err for testing.
+func NewStatusResult(val string, err error) *StatusCmd {
+	var cmd StatusCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewIntResult returns an IntCmd initialised with val and err for testing.
+func NewIntResult(val int64, err error) *IntCmd {
+	var cmd IntCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewDurationResult returns a DurationCmd initialised with val and err for testing.
+func NewDurationResult(val time.Duration, err error) *DurationCmd {
+	var cmd DurationCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewBoolResult returns a BoolCmd initialised with val and err for testing.
+func NewBoolResult(val bool, err error) *BoolCmd {
+	var cmd BoolCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewStringResult returns a StringCmd initialised with val and err for testing.
+func NewStringResult(val string, err error) *StringCmd {
+	var cmd StringCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewFloatResult returns a FloatCmd initialised with val and err for testing.
+func NewFloatResult(val float64, err error) *FloatCmd {
+	var cmd FloatCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewStringSliceResult returns a StringSliceCmd initialised with val and err for testing.
+func NewStringSliceResult(val []string, err error) *StringSliceCmd {
+	var cmd StringSliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewBoolSliceResult returns a BoolSliceCmd initialised with val and err for testing.
+func NewBoolSliceResult(val []bool, err error) *BoolSliceCmd {
+	var cmd BoolSliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewStringStringMapResult returns a StringStringMapCmd initialised with val and err for testing.
+func NewStringStringMapResult(val map[string]string, err error) *StringStringMapCmd {
+	var cmd StringStringMapCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewStringIntMapCmdResult returns a StringIntMapCmd initialised with val and err for testing.
+func NewStringIntMapCmdResult(val map[string]int64, err error) *StringIntMapCmd {
+	var cmd StringIntMapCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewTimeCmdResult returns a TimeCmd initialised with val and err for testing.
+func NewTimeCmdResult(val time.Time, err error) *TimeCmd {
+	var cmd TimeCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewZSliceCmdResult returns a ZSliceCmd initialised with val and err for testing.
+func NewZSliceCmdResult(val []Z, err error) *ZSliceCmd {
+	var cmd ZSliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewZWithKeyCmdResult returns a NewZWithKeyCmd initialised with val and err for testing.
+func NewZWithKeyCmdResult(val *ZWithKey, err error) *ZWithKeyCmd {
+	var cmd ZWithKeyCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewScanCmdResult returns a ScanCmd initialised with val and err for testing.
+func NewScanCmdResult(keys []string, cursor uint64, err error) *ScanCmd {
+	var cmd ScanCmd
+	cmd.page = keys
+	cmd.cursor = cursor
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewClusterSlotsCmdResult returns a ClusterSlotsCmd initialised with val and err for testing.
+func NewClusterSlotsCmdResult(val []ClusterSlot, err error) *ClusterSlotsCmd {
+	var cmd ClusterSlotsCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewGeoLocationCmdResult returns a GeoLocationCmd initialised with val and err for testing.
+func NewGeoLocationCmdResult(val []GeoLocation, err error) *GeoLocationCmd {
+	var cmd GeoLocationCmd
+	cmd.locations = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewGeoPosCmdResult returns a GeoPosCmd initialised with val and err for testing.
+func NewGeoPosCmdResult(val []*GeoPos, err error) *GeoPosCmd {
+	var cmd GeoPosCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewCommandsInfoCmdResult returns a CommandsInfoCmd initialised with val and err for testing.
+func NewCommandsInfoCmdResult(val map[string]*CommandInfo, err error) *CommandsInfoCmd {
+	var cmd CommandsInfoCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewXMessageSliceCmdResult returns a XMessageSliceCmd initialised with val and err for testing.
+func NewXMessageSliceCmdResult(val []XMessage, err error) *XMessageSliceCmd {
+	var cmd XMessageSliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
+
+// NewXStreamSliceCmdResult returns a XStreamSliceCmd initialised with val and err for testing.
+func NewXStreamSliceCmdResult(val []XStream, err error) *XStreamSliceCmd {
+	var cmd XStreamSliceCmd
+	cmd.val = val
+	cmd.SetErr(err)
+	return &cmd
+}
diff --git a/vendor/github.com/go-redis/redis/v8/ring.go b/vendor/github.com/go-redis/redis/v8/ring.go
new file mode 100644
index 0000000..34d05f3
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/ring.go
@@ -0,0 +1,731 @@
+package redis
+
+import (
+	"context"
+	"crypto/tls"
+	"errors"
+	"fmt"
+	"net"
+	"strconv"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	"github.com/cespare/xxhash/v2"
+	"github.com/dgryski/go-rendezvous"
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/hashtag"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/rand"
+)
+
+var errRingShardsDown = errors.New("redis: all ring shards are down")
+
+//------------------------------------------------------------------------------
+
+type ConsistentHash interface {
+	Get(string) string
+}
+
+type rendezvousWrapper struct {
+	*rendezvous.Rendezvous
+}
+
+func (w rendezvousWrapper) Get(key string) string {
+	return w.Lookup(key)
+}
+
+func newRendezvous(shards []string) ConsistentHash {
+	return rendezvousWrapper{rendezvous.New(shards, xxhash.Sum64String)}
+}
+
+//------------------------------------------------------------------------------
+
+// RingOptions are used to configure a ring client and should be
+// passed to NewRing.
+type RingOptions struct {
+	// Map of name => host:port addresses of ring shards.
+	Addrs map[string]string
+
+	// NewClient creates a shard client with provided name and options.
+	NewClient func(name string, opt *Options) *Client
+
+	// Frequency of PING commands sent to check shards availability.
+	// Shard is considered down after 3 subsequent failed checks.
+	HeartbeatFrequency time.Duration
+
+	// NewConsistentHash returns a consistent hash that is used
+	// to distribute keys across the shards.
+	//
+	// See https://medium.com/@dgryski/consistent-hashing-algorithmic-tradeoffs-ef6b8e2fcae8
+	// for consistent hashing algorithmic tradeoffs.
+	NewConsistentHash func(shards []string) ConsistentHash
+
+	// Following options are copied from Options struct.
+
+	Dialer    func(ctx context.Context, network, addr string) (net.Conn, error)
+	OnConnect func(ctx context.Context, cn *Conn) error
+
+	Username string
+	Password string
+	DB       int
+
+	MaxRetries      int
+	MinRetryBackoff time.Duration
+	MaxRetryBackoff time.Duration
+
+	DialTimeout  time.Duration
+	ReadTimeout  time.Duration
+	WriteTimeout time.Duration
+
+	PoolSize           int
+	MinIdleConns       int
+	MaxConnAge         time.Duration
+	PoolTimeout        time.Duration
+	IdleTimeout        time.Duration
+	IdleCheckFrequency time.Duration
+
+	TLSConfig *tls.Config
+	Limiter   Limiter
+}
+
+func (opt *RingOptions) init() {
+	if opt.NewClient == nil {
+		opt.NewClient = func(name string, opt *Options) *Client {
+			return NewClient(opt)
+		}
+	}
+
+	if opt.HeartbeatFrequency == 0 {
+		opt.HeartbeatFrequency = 500 * time.Millisecond
+	}
+
+	if opt.NewConsistentHash == nil {
+		opt.NewConsistentHash = newRendezvous
+	}
+
+	if opt.MaxRetries == -1 {
+		opt.MaxRetries = 0
+	} else if opt.MaxRetries == 0 {
+		opt.MaxRetries = 3
+	}
+	switch opt.MinRetryBackoff {
+	case -1:
+		opt.MinRetryBackoff = 0
+	case 0:
+		opt.MinRetryBackoff = 8 * time.Millisecond
+	}
+	switch opt.MaxRetryBackoff {
+	case -1:
+		opt.MaxRetryBackoff = 0
+	case 0:
+		opt.MaxRetryBackoff = 512 * time.Millisecond
+	}
+}
+
+func (opt *RingOptions) clientOptions() *Options {
+	return &Options{
+		Dialer:    opt.Dialer,
+		OnConnect: opt.OnConnect,
+
+		Username: opt.Username,
+		Password: opt.Password,
+		DB:       opt.DB,
+
+		MaxRetries: -1,
+
+		DialTimeout:  opt.DialTimeout,
+		ReadTimeout:  opt.ReadTimeout,
+		WriteTimeout: opt.WriteTimeout,
+
+		PoolSize:           opt.PoolSize,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: opt.IdleCheckFrequency,
+
+		TLSConfig: opt.TLSConfig,
+		Limiter:   opt.Limiter,
+	}
+}
+
+//------------------------------------------------------------------------------
+
+type ringShard struct {
+	Client *Client
+	down   int32
+}
+
+func newRingShard(opt *RingOptions, name, addr string) *ringShard {
+	clopt := opt.clientOptions()
+	clopt.Addr = addr
+
+	return &ringShard{
+		Client: opt.NewClient(name, clopt),
+	}
+}
+
+func (shard *ringShard) String() string {
+	var state string
+	if shard.IsUp() {
+		state = "up"
+	} else {
+		state = "down"
+	}
+	return fmt.Sprintf("%s is %s", shard.Client, state)
+}
+
+func (shard *ringShard) IsDown() bool {
+	const threshold = 3
+	return atomic.LoadInt32(&shard.down) >= threshold
+}
+
+func (shard *ringShard) IsUp() bool {
+	return !shard.IsDown()
+}
+
+// Vote votes to set shard state and returns true if state was changed.
+func (shard *ringShard) Vote(up bool) bool {
+	if up {
+		changed := shard.IsDown()
+		atomic.StoreInt32(&shard.down, 0)
+		return changed
+	}
+
+	if shard.IsDown() {
+		return false
+	}
+
+	atomic.AddInt32(&shard.down, 1)
+	return shard.IsDown()
+}
+
+//------------------------------------------------------------------------------
+
+type ringShards struct {
+	opt *RingOptions
+
+	mu       sync.RWMutex
+	hash     ConsistentHash
+	shards   map[string]*ringShard // read only
+	list     []*ringShard          // read only
+	numShard int
+	closed   bool
+}
+
+func newRingShards(opt *RingOptions) *ringShards {
+	shards := make(map[string]*ringShard, len(opt.Addrs))
+	list := make([]*ringShard, 0, len(shards))
+
+	for name, addr := range opt.Addrs {
+		shard := newRingShard(opt, name, addr)
+		shards[name] = shard
+
+		list = append(list, shard)
+	}
+
+	c := &ringShards{
+		opt: opt,
+
+		shards: shards,
+		list:   list,
+	}
+	c.rebalance()
+
+	return c
+}
+
+func (c *ringShards) List() []*ringShard {
+	var list []*ringShard
+
+	c.mu.RLock()
+	if !c.closed {
+		list = c.list
+	}
+	c.mu.RUnlock()
+
+	return list
+}
+
+func (c *ringShards) Hash(key string) string {
+	key = hashtag.Key(key)
+
+	var hash string
+
+	c.mu.RLock()
+	if c.numShard > 0 {
+		hash = c.hash.Get(key)
+	}
+	c.mu.RUnlock()
+
+	return hash
+}
+
+func (c *ringShards) GetByKey(key string) (*ringShard, error) {
+	key = hashtag.Key(key)
+
+	c.mu.RLock()
+
+	if c.closed {
+		c.mu.RUnlock()
+		return nil, pool.ErrClosed
+	}
+
+	if c.numShard == 0 {
+		c.mu.RUnlock()
+		return nil, errRingShardsDown
+	}
+
+	hash := c.hash.Get(key)
+	if hash == "" {
+		c.mu.RUnlock()
+		return nil, errRingShardsDown
+	}
+
+	shard := c.shards[hash]
+	c.mu.RUnlock()
+
+	return shard, nil
+}
+
+func (c *ringShards) GetByName(shardName string) (*ringShard, error) {
+	if shardName == "" {
+		return c.Random()
+	}
+
+	c.mu.RLock()
+	shard := c.shards[shardName]
+	c.mu.RUnlock()
+	return shard, nil
+}
+
+func (c *ringShards) Random() (*ringShard, error) {
+	return c.GetByKey(strconv.Itoa(rand.Int()))
+}
+
+// heartbeat monitors state of each shard in the ring.
+func (c *ringShards) Heartbeat(frequency time.Duration) {
+	ticker := time.NewTicker(frequency)
+	defer ticker.Stop()
+
+	ctx := context.Background()
+	for range ticker.C {
+		var rebalance bool
+
+		for _, shard := range c.List() {
+			err := shard.Client.Ping(ctx).Err()
+			isUp := err == nil || err == pool.ErrPoolTimeout
+			if shard.Vote(isUp) {
+				internal.Logger.Printf(context.Background(), "ring shard state changed: %s", shard)
+				rebalance = true
+			}
+		}
+
+		if rebalance {
+			c.rebalance()
+		}
+	}
+}
+
+// rebalance removes dead shards from the Ring.
+func (c *ringShards) rebalance() {
+	c.mu.RLock()
+	shards := c.shards
+	c.mu.RUnlock()
+
+	liveShards := make([]string, 0, len(shards))
+
+	for name, shard := range shards {
+		if shard.IsUp() {
+			liveShards = append(liveShards, name)
+		}
+	}
+
+	hash := c.opt.NewConsistentHash(liveShards)
+
+	c.mu.Lock()
+	c.hash = hash
+	c.numShard = len(liveShards)
+	c.mu.Unlock()
+}
+
+func (c *ringShards) Len() int {
+	c.mu.RLock()
+	l := c.numShard
+	c.mu.RUnlock()
+	return l
+}
+
+func (c *ringShards) Close() error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.closed {
+		return nil
+	}
+	c.closed = true
+
+	var firstErr error
+	for _, shard := range c.shards {
+		if err := shard.Client.Close(); err != nil && firstErr == nil {
+			firstErr = err
+		}
+	}
+	c.hash = nil
+	c.shards = nil
+	c.list = nil
+
+	return firstErr
+}
+
+//------------------------------------------------------------------------------
+
+type ring struct {
+	opt           *RingOptions
+	shards        *ringShards
+	cmdsInfoCache *cmdsInfoCache //nolint:structcheck
+}
+
+// Ring is a Redis client that uses consistent hashing to distribute
+// keys across multiple Redis servers (shards). It's safe for
+// concurrent use by multiple goroutines.
+//
+// Ring monitors the state of each shard and removes dead shards from
+// the ring. When a shard comes online it is added back to the ring. This
+// gives you maximum availability and partition tolerance, but no
+// consistency between different shards or even clients. Each client
+// uses shards that are available to the client and does not do any
+// coordination when shard state is changed.
+//
+// Ring should be used when you need multiple Redis servers for caching
+// and can tolerate losing data when one of the servers dies.
+// Otherwise you should use Redis Cluster.
+type Ring struct {
+	*ring
+	cmdable
+	hooks
+	ctx context.Context
+}
+
+func NewRing(opt *RingOptions) *Ring {
+	opt.init()
+
+	ring := Ring{
+		ring: &ring{
+			opt:    opt,
+			shards: newRingShards(opt),
+		},
+		ctx: context.Background(),
+	}
+
+	ring.cmdsInfoCache = newCmdsInfoCache(ring.cmdsInfo)
+	ring.cmdable = ring.Process
+
+	go ring.shards.Heartbeat(opt.HeartbeatFrequency)
+
+	return &ring
+}
+
+func (c *Ring) Context() context.Context {
+	return c.ctx
+}
+
+func (c *Ring) WithContext(ctx context.Context) *Ring {
+	if ctx == nil {
+		panic("nil context")
+	}
+	clone := *c
+	clone.cmdable = clone.Process
+	clone.hooks.lock()
+	clone.ctx = ctx
+	return &clone
+}
+
+// Do creates a Cmd from the args and processes the cmd.
+func (c *Ring) Do(ctx context.Context, args ...interface{}) *Cmd {
+	cmd := NewCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+func (c *Ring) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.process)
+}
+
+// Options returns read-only Options that were used to create the client.
+func (c *Ring) Options() *RingOptions {
+	return c.opt
+}
+
+func (c *Ring) retryBackoff(attempt int) time.Duration {
+	return internal.RetryBackoff(attempt, c.opt.MinRetryBackoff, c.opt.MaxRetryBackoff)
+}
+
+// PoolStats returns accumulated connection pool stats.
+func (c *Ring) PoolStats() *PoolStats {
+	shards := c.shards.List()
+	var acc PoolStats
+	for _, shard := range shards {
+		s := shard.Client.connPool.Stats()
+		acc.Hits += s.Hits
+		acc.Misses += s.Misses
+		acc.Timeouts += s.Timeouts
+		acc.TotalConns += s.TotalConns
+		acc.IdleConns += s.IdleConns
+	}
+	return &acc
+}
+
+// Len returns the current number of shards in the ring.
+func (c *Ring) Len() int {
+	return c.shards.Len()
+}
+
+// Subscribe subscribes the client to the specified channels.
+func (c *Ring) Subscribe(ctx context.Context, channels ...string) *PubSub {
+	if len(channels) == 0 {
+		panic("at least one channel is required")
+	}
+
+	shard, err := c.shards.GetByKey(channels[0])
+	if err != nil {
+		// TODO: return PubSub with sticky error
+		panic(err)
+	}
+	return shard.Client.Subscribe(ctx, channels...)
+}
+
+// PSubscribe subscribes the client to the given patterns.
+func (c *Ring) PSubscribe(ctx context.Context, channels ...string) *PubSub {
+	if len(channels) == 0 {
+		panic("at least one channel is required")
+	}
+
+	shard, err := c.shards.GetByKey(channels[0])
+	if err != nil {
+		// TODO: return PubSub with sticky error
+		panic(err)
+	}
+	return shard.Client.PSubscribe(ctx, channels...)
+}
+
+// ForEachShard concurrently calls the fn on each live shard in the ring.
+// It returns the first error if any.
+func (c *Ring) ForEachShard(
+	ctx context.Context,
+	fn func(ctx context.Context, client *Client) error,
+) error {
+	shards := c.shards.List()
+	var wg sync.WaitGroup
+	errCh := make(chan error, 1)
+	for _, shard := range shards {
+		if shard.IsDown() {
+			continue
+		}
+
+		wg.Add(1)
+		go func(shard *ringShard) {
+			defer wg.Done()
+			err := fn(ctx, shard.Client)
+			if err != nil {
+				select {
+				case errCh <- err:
+				default:
+				}
+			}
+		}(shard)
+	}
+	wg.Wait()
+
+	select {
+	case err := <-errCh:
+		return err
+	default:
+		return nil
+	}
+}
+
+func (c *Ring) cmdsInfo(ctx context.Context) (map[string]*CommandInfo, error) {
+	shards := c.shards.List()
+	var firstErr error
+	for _, shard := range shards {
+		cmdsInfo, err := shard.Client.Command(ctx).Result()
+		if err == nil {
+			return cmdsInfo, nil
+		}
+		if firstErr == nil {
+			firstErr = err
+		}
+	}
+	if firstErr == nil {
+		return nil, errRingShardsDown
+	}
+	return nil, firstErr
+}
+
+func (c *Ring) cmdInfo(ctx context.Context, name string) *CommandInfo {
+	cmdsInfo, err := c.cmdsInfoCache.Get(ctx)
+	if err != nil {
+		return nil
+	}
+	info := cmdsInfo[name]
+	if info == nil {
+		internal.Logger.Printf(c.Context(), "info for cmd=%s not found", name)
+	}
+	return info
+}
+
+func (c *Ring) cmdShard(ctx context.Context, cmd Cmder) (*ringShard, error) {
+	cmdInfo := c.cmdInfo(ctx, cmd.Name())
+	pos := cmdFirstKeyPos(cmd, cmdInfo)
+	if pos == 0 {
+		return c.shards.Random()
+	}
+	firstKey := cmd.stringArg(pos)
+	return c.shards.GetByKey(firstKey)
+}
+
+func (c *Ring) process(ctx context.Context, cmd Cmder) error {
+	var lastErr error
+	for attempt := 0; attempt <= c.opt.MaxRetries; attempt++ {
+		if attempt > 0 {
+			if err := internal.Sleep(ctx, c.retryBackoff(attempt)); err != nil {
+				return err
+			}
+		}
+
+		shard, err := c.cmdShard(ctx, cmd)
+		if err != nil {
+			return err
+		}
+
+		lastErr = shard.Client.Process(ctx, cmd)
+		if lastErr == nil || !shouldRetry(lastErr, cmd.readTimeout() == nil) {
+			return lastErr
+		}
+	}
+	return lastErr
+}
+
+func (c *Ring) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipeline().Pipelined(ctx, fn)
+}
+
+func (c *Ring) Pipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *Ring) processPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, func(ctx context.Context, cmds []Cmder) error {
+		return c.generalProcessPipeline(ctx, cmds, false)
+	})
+}
+
+func (c *Ring) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.TxPipeline().Pipelined(ctx, fn)
+}
+
+func (c *Ring) TxPipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx:  c.ctx,
+		exec: c.processTxPipeline,
+	}
+	pipe.init()
+	return &pipe
+}
+
+func (c *Ring) processTxPipeline(ctx context.Context, cmds []Cmder) error {
+	return c.hooks.processPipeline(ctx, cmds, func(ctx context.Context, cmds []Cmder) error {
+		return c.generalProcessPipeline(ctx, cmds, true)
+	})
+}
+
+func (c *Ring) generalProcessPipeline(
+	ctx context.Context, cmds []Cmder, tx bool,
+) error {
+	cmdsMap := make(map[string][]Cmder)
+	for _, cmd := range cmds {
+		cmdInfo := c.cmdInfo(ctx, cmd.Name())
+		hash := cmd.stringArg(cmdFirstKeyPos(cmd, cmdInfo))
+		if hash != "" {
+			hash = c.shards.Hash(hash)
+		}
+		cmdsMap[hash] = append(cmdsMap[hash], cmd)
+	}
+
+	var wg sync.WaitGroup
+	for hash, cmds := range cmdsMap {
+		wg.Add(1)
+		go func(hash string, cmds []Cmder) {
+			defer wg.Done()
+
+			_ = c.processShardPipeline(ctx, hash, cmds, tx)
+		}(hash, cmds)
+	}
+
+	wg.Wait()
+	return cmdsFirstErr(cmds)
+}
+
+func (c *Ring) processShardPipeline(
+	ctx context.Context, hash string, cmds []Cmder, tx bool,
+) error {
+	// TODO: retry?
+	shard, err := c.shards.GetByName(hash)
+	if err != nil {
+		setCmdsErr(cmds, err)
+		return err
+	}
+
+	if tx {
+		return shard.Client.processTxPipeline(ctx, cmds)
+	}
+	return shard.Client.processPipeline(ctx, cmds)
+}
+
+func (c *Ring) Watch(ctx context.Context, fn func(*Tx) error, keys ...string) error {
+	if len(keys) == 0 {
+		return fmt.Errorf("redis: Watch requires at least one key")
+	}
+
+	var shards []*ringShard
+	for _, key := range keys {
+		if key != "" {
+			shard, err := c.shards.GetByKey(hashtag.Key(key))
+			if err != nil {
+				return err
+			}
+
+			shards = append(shards, shard)
+		}
+	}
+
+	if len(shards) == 0 {
+		return fmt.Errorf("redis: Watch requires at least one shard")
+	}
+
+	if len(shards) > 1 {
+		for _, shard := range shards[1:] {
+			if shard.Client != shards[0].Client {
+				err := fmt.Errorf("redis: Watch requires all keys to be in the same shard")
+				return err
+			}
+		}
+	}
+
+	return shards[0].Client.Watch(ctx, fn, keys...)
+}
+
+// Close closes the ring client, releasing any open resources.
+//
+// It is rare to Close a Ring, as the Ring is meant to be long-lived
+// and shared between many goroutines.
+func (c *Ring) Close() error {
+	return c.shards.Close()
+}
diff --git a/vendor/github.com/go-redis/redis/v8/script.go b/vendor/github.com/go-redis/redis/v8/script.go
new file mode 100644
index 0000000..07ed482
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/script.go
@@ -0,0 +1,65 @@
+package redis
+
+import (
+	"context"
+	"crypto/sha1"
+	"encoding/hex"
+	"io"
+	"strings"
+)
+
+type scripter interface {
+	Eval(ctx context.Context, script string, keys []string, args ...interface{}) *Cmd
+	EvalSha(ctx context.Context, sha1 string, keys []string, args ...interface{}) *Cmd
+	ScriptExists(ctx context.Context, hashes ...string) *BoolSliceCmd
+	ScriptLoad(ctx context.Context, script string) *StringCmd
+}
+
+var (
+	_ scripter = (*Client)(nil)
+	_ scripter = (*Ring)(nil)
+	_ scripter = (*ClusterClient)(nil)
+)
+
+type Script struct {
+	src, hash string
+}
+
+func NewScript(src string) *Script {
+	h := sha1.New()
+	_, _ = io.WriteString(h, src)
+	return &Script{
+		src:  src,
+		hash: hex.EncodeToString(h.Sum(nil)),
+	}
+}
+
+func (s *Script) Hash() string {
+	return s.hash
+}
+
+func (s *Script) Load(ctx context.Context, c scripter) *StringCmd {
+	return c.ScriptLoad(ctx, s.src)
+}
+
+func (s *Script) Exists(ctx context.Context, c scripter) *BoolSliceCmd {
+	return c.ScriptExists(ctx, s.hash)
+}
+
+func (s *Script) Eval(ctx context.Context, c scripter, keys []string, args ...interface{}) *Cmd {
+	return c.Eval(ctx, s.src, keys, args...)
+}
+
+func (s *Script) EvalSha(ctx context.Context, c scripter, keys []string, args ...interface{}) *Cmd {
+	return c.EvalSha(ctx, s.hash, keys, args...)
+}
+
+// Run optimistically uses EVALSHA to run the script. If script does not exist
+// it is retried using EVAL.
+func (s *Script) Run(ctx context.Context, c scripter, keys []string, args ...interface{}) *Cmd {
+	r := s.EvalSha(ctx, c, keys, args...)
+	if err := r.Err(); err != nil && strings.HasPrefix(err.Error(), "NOSCRIPT ") {
+		return s.Eval(ctx, c, keys, args...)
+	}
+	return r
+}
diff --git a/vendor/github.com/go-redis/redis/v8/sentinel.go b/vendor/github.com/go-redis/redis/v8/sentinel.go
new file mode 100644
index 0000000..7db9843
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/sentinel.go
@@ -0,0 +1,731 @@
+package redis
+
+import (
+	"context"
+	"crypto/tls"
+	"errors"
+	"net"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/go-redis/redis/v8/internal"
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/rand"
+)
+
+//------------------------------------------------------------------------------
+
+// FailoverOptions are used to configure a failover client and should
+// be passed to NewFailoverClient.
+type FailoverOptions struct {
+	// The master name.
+	MasterName string
+	// A seed list of host:port addresses of sentinel nodes.
+	SentinelAddrs []string
+	// Sentinel password from "requirepass <password>" (if enabled) in Sentinel configuration
+	SentinelPassword string
+
+	// Allows routing read-only commands to the closest master or slave node.
+	// This option only works with NewFailoverClusterClient.
+	RouteByLatency bool
+	// Allows routing read-only commands to the random master or slave node.
+	// This option only works with NewFailoverClusterClient.
+	RouteRandomly bool
+
+	// Route all commands to slave read-only nodes.
+	SlaveOnly bool
+
+	// Following options are copied from Options struct.
+
+	Dialer    func(ctx context.Context, network, addr string) (net.Conn, error)
+	OnConnect func(ctx context.Context, cn *Conn) error
+
+	Username string
+	Password string
+	DB       int
+
+	MaxRetries      int
+	MinRetryBackoff time.Duration
+	MaxRetryBackoff time.Duration
+
+	DialTimeout  time.Duration
+	ReadTimeout  time.Duration
+	WriteTimeout time.Duration
+
+	PoolSize           int
+	MinIdleConns       int
+	MaxConnAge         time.Duration
+	PoolTimeout        time.Duration
+	IdleTimeout        time.Duration
+	IdleCheckFrequency time.Duration
+
+	TLSConfig *tls.Config
+}
+
+func (opt *FailoverOptions) clientOptions() *Options {
+	return &Options{
+		Addr: "FailoverClient",
+
+		Dialer:    opt.Dialer,
+		OnConnect: opt.OnConnect,
+
+		DB:       opt.DB,
+		Username: opt.Username,
+		Password: opt.Password,
+
+		MaxRetries:      opt.MaxRetries,
+		MinRetryBackoff: opt.MinRetryBackoff,
+		MaxRetryBackoff: opt.MaxRetryBackoff,
+
+		DialTimeout:  opt.DialTimeout,
+		ReadTimeout:  opt.ReadTimeout,
+		WriteTimeout: opt.WriteTimeout,
+
+		PoolSize:           opt.PoolSize,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: opt.IdleCheckFrequency,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+
+		TLSConfig: opt.TLSConfig,
+	}
+}
+
+func (opt *FailoverOptions) sentinelOptions(addr string) *Options {
+	return &Options{
+		Addr: addr,
+
+		Dialer:    opt.Dialer,
+		OnConnect: opt.OnConnect,
+
+		DB:       0,
+		Password: opt.SentinelPassword,
+
+		MaxRetries:      opt.MaxRetries,
+		MinRetryBackoff: opt.MinRetryBackoff,
+		MaxRetryBackoff: opt.MaxRetryBackoff,
+
+		DialTimeout:  opt.DialTimeout,
+		ReadTimeout:  opt.ReadTimeout,
+		WriteTimeout: opt.WriteTimeout,
+
+		PoolSize:           opt.PoolSize,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: opt.IdleCheckFrequency,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+
+		TLSConfig: opt.TLSConfig,
+	}
+}
+
+func (opt *FailoverOptions) clusterOptions() *ClusterOptions {
+	return &ClusterOptions{
+		Dialer:    opt.Dialer,
+		OnConnect: opt.OnConnect,
+
+		Username: opt.Username,
+		Password: opt.Password,
+
+		MaxRedirects: opt.MaxRetries,
+
+		RouteByLatency: opt.RouteByLatency,
+		RouteRandomly:  opt.RouteRandomly,
+
+		MinRetryBackoff: opt.MinRetryBackoff,
+		MaxRetryBackoff: opt.MaxRetryBackoff,
+
+		DialTimeout:  opt.DialTimeout,
+		ReadTimeout:  opt.ReadTimeout,
+		WriteTimeout: opt.WriteTimeout,
+
+		PoolSize:           opt.PoolSize,
+		PoolTimeout:        opt.PoolTimeout,
+		IdleTimeout:        opt.IdleTimeout,
+		IdleCheckFrequency: opt.IdleCheckFrequency,
+		MinIdleConns:       opt.MinIdleConns,
+		MaxConnAge:         opt.MaxConnAge,
+
+		TLSConfig: opt.TLSConfig,
+	}
+}
+
+// NewFailoverClient returns a Redis client that uses Redis Sentinel
+// for automatic failover. It's safe for concurrent use by multiple
+// goroutines.
+func NewFailoverClient(failoverOpt *FailoverOptions) *Client {
+	if failoverOpt.RouteByLatency {
+		panic("to route commands by latency, use NewFailoverClusterClient")
+	}
+	if failoverOpt.RouteRandomly {
+		panic("to route commands randomly, use NewFailoverClusterClient")
+	}
+
+	sentinelAddrs := make([]string, len(failoverOpt.SentinelAddrs))
+	copy(sentinelAddrs, failoverOpt.SentinelAddrs)
+
+	failover := &sentinelFailover{
+		opt:           failoverOpt,
+		sentinelAddrs: sentinelAddrs,
+	}
+
+	opt := failoverOpt.clientOptions()
+	opt.Dialer = masterSlaveDialer(failover)
+	opt.init()
+
+	connPool := newConnPool(opt)
+	failover.onFailover = func(ctx context.Context, addr string) {
+		_ = connPool.Filter(func(cn *pool.Conn) bool {
+			return cn.RemoteAddr().String() != addr
+		})
+	}
+
+	c := Client{
+		baseClient: newBaseClient(opt, connPool),
+		ctx:        context.Background(),
+	}
+	c.cmdable = c.Process
+	c.onClose = failover.Close
+
+	return &c
+}
+
+func masterSlaveDialer(
+	failover *sentinelFailover,
+) func(ctx context.Context, network, addr string) (net.Conn, error) {
+	return func(ctx context.Context, network, _ string) (net.Conn, error) {
+		var addr string
+		var err error
+
+		if failover.opt.SlaveOnly {
+			addr, err = failover.RandomSlaveAddr(ctx)
+		} else {
+			addr, err = failover.MasterAddr(ctx)
+			if err == nil {
+				failover.trySwitchMaster(ctx, addr)
+			}
+		}
+
+		if err != nil {
+			return nil, err
+		}
+		if failover.opt.Dialer != nil {
+			return failover.opt.Dialer(ctx, network, addr)
+		}
+		return net.DialTimeout("tcp", addr, failover.opt.DialTimeout)
+	}
+}
+
+//------------------------------------------------------------------------------
+
+// SentinelClient is a client for a Redis Sentinel.
+type SentinelClient struct {
+	*baseClient
+	hooks
+	ctx context.Context
+}
+
+func NewSentinelClient(opt *Options) *SentinelClient {
+	opt.init()
+	c := &SentinelClient{
+		baseClient: &baseClient{
+			opt:      opt,
+			connPool: newConnPool(opt),
+		},
+		ctx: context.Background(),
+	}
+	return c
+}
+
+func (c *SentinelClient) Context() context.Context {
+	return c.ctx
+}
+
+func (c *SentinelClient) WithContext(ctx context.Context) *SentinelClient {
+	if ctx == nil {
+		panic("nil context")
+	}
+	clone := *c
+	clone.ctx = ctx
+	return &clone
+}
+
+func (c *SentinelClient) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.baseClient.process)
+}
+
+func (c *SentinelClient) pubSub() *PubSub {
+	pubsub := &PubSub{
+		opt: c.opt,
+
+		newConn: func(ctx context.Context, channels []string) (*pool.Conn, error) {
+			return c.newConn(ctx)
+		},
+		closeConn: c.connPool.CloseConn,
+	}
+	pubsub.init()
+	return pubsub
+}
+
+// Ping is used to test if a connection is still alive, or to
+// measure latency.
+func (c *SentinelClient) Ping(ctx context.Context) *StringCmd {
+	cmd := NewStringCmd(ctx, "ping")
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Subscribe subscribes the client to the specified channels.
+// Channels can be omitted to create empty subscription.
+func (c *SentinelClient) Subscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.Subscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+// PSubscribe subscribes the client to the given patterns.
+// Patterns can be omitted to create empty subscription.
+func (c *SentinelClient) PSubscribe(ctx context.Context, channels ...string) *PubSub {
+	pubsub := c.pubSub()
+	if len(channels) > 0 {
+		_ = pubsub.PSubscribe(ctx, channels...)
+	}
+	return pubsub
+}
+
+func (c *SentinelClient) GetMasterAddrByName(ctx context.Context, name string) *StringSliceCmd {
+	cmd := NewStringSliceCmd(ctx, "sentinel", "get-master-addr-by-name", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+func (c *SentinelClient) Sentinels(ctx context.Context, name string) *SliceCmd {
+	cmd := NewSliceCmd(ctx, "sentinel", "sentinels", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Failover forces a failover as if the master was not reachable, and without
+// asking for agreement to other Sentinels.
+func (c *SentinelClient) Failover(ctx context.Context, name string) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "sentinel", "failover", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Reset resets all the masters with matching name. The pattern argument is a
+// glob-style pattern. The reset process clears any previous state in a master
+// (including a failover in progress), and removes every slave and sentinel
+// already discovered and associated with the master.
+func (c *SentinelClient) Reset(ctx context.Context, pattern string) *IntCmd {
+	cmd := NewIntCmd(ctx, "sentinel", "reset", pattern)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// FlushConfig forces Sentinel to rewrite its configuration on disk, including
+// the current Sentinel state.
+func (c *SentinelClient) FlushConfig(ctx context.Context) *StatusCmd {
+	cmd := NewStatusCmd(ctx, "sentinel", "flushconfig")
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Master shows the state and info of the specified master.
+func (c *SentinelClient) Master(ctx context.Context, name string) *StringStringMapCmd {
+	cmd := NewStringStringMapCmd(ctx, "sentinel", "master", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Masters shows a list of monitored masters and their state.
+func (c *SentinelClient) Masters(ctx context.Context) *SliceCmd {
+	cmd := NewSliceCmd(ctx, "sentinel", "masters")
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Slaves shows a list of slaves for the specified master and their state.
+func (c *SentinelClient) Slaves(ctx context.Context, name string) *SliceCmd {
+	cmd := NewSliceCmd(ctx, "sentinel", "slaves", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// CkQuorum checks if the current Sentinel configuration is able to reach the
+// quorum needed to failover a master, and the majority needed to authorize the
+// failover. This command should be used in monitoring systems to check if a
+// Sentinel deployment is ok.
+func (c *SentinelClient) CkQuorum(ctx context.Context, name string) *StringCmd {
+	cmd := NewStringCmd(ctx, "sentinel", "ckquorum", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Monitor tells the Sentinel to start monitoring a new master with the specified
+// name, ip, port, and quorum.
+func (c *SentinelClient) Monitor(ctx context.Context, name, ip, port, quorum string) *StringCmd {
+	cmd := NewStringCmd(ctx, "sentinel", "monitor", name, ip, port, quorum)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Set is used in order to change configuration parameters of a specific master.
+func (c *SentinelClient) Set(ctx context.Context, name, option, value string) *StringCmd {
+	cmd := NewStringCmd(ctx, "sentinel", "set", name, option, value)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Remove is used in order to remove the specified master: the master will no
+// longer be monitored, and will totally be removed from the internal state of
+// the Sentinel.
+func (c *SentinelClient) Remove(ctx context.Context, name string) *StringCmd {
+	cmd := NewStringCmd(ctx, "sentinel", "remove", name)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+//------------------------------------------------------------------------------
+
+type sentinelFailover struct {
+	opt *FailoverOptions
+
+	sentinelAddrs []string
+
+	onFailover func(ctx context.Context, addr string)
+	onUpdate   func(ctx context.Context)
+
+	mu          sync.RWMutex
+	_masterAddr string
+	sentinel    *SentinelClient
+	pubsub      *PubSub
+}
+
+func (c *sentinelFailover) Close() error {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	if c.sentinel != nil {
+		return c.closeSentinel()
+	}
+	return nil
+}
+
+func (c *sentinelFailover) closeSentinel() error {
+	firstErr := c.pubsub.Close()
+	c.pubsub = nil
+
+	err := c.sentinel.Close()
+	if err != nil && firstErr == nil {
+		firstErr = err
+	}
+	c.sentinel = nil
+
+	return firstErr
+}
+
+func (c *sentinelFailover) RandomSlaveAddr(ctx context.Context) (string, error) {
+	addresses, err := c.slaveAddrs(ctx)
+	if err != nil {
+		return "", err
+	}
+	if len(addresses) == 0 {
+		return c.MasterAddr(ctx)
+	}
+	return addresses[rand.Intn(len(addresses))], nil
+}
+
+func (c *sentinelFailover) MasterAddr(ctx context.Context) (string, error) {
+	c.mu.RLock()
+	sentinel := c.sentinel
+	c.mu.RUnlock()
+
+	if sentinel != nil {
+		addr := c.getMasterAddr(ctx, sentinel)
+		if addr != "" {
+			return addr, nil
+		}
+	}
+
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.sentinel != nil {
+		addr := c.getMasterAddr(ctx, c.sentinel)
+		if addr != "" {
+			return addr, nil
+		}
+		_ = c.closeSentinel()
+	}
+
+	for i, sentinelAddr := range c.sentinelAddrs {
+		sentinel := NewSentinelClient(c.opt.sentinelOptions(sentinelAddr))
+
+		masterAddr, err := sentinel.GetMasterAddrByName(ctx, c.opt.MasterName).Result()
+		if err != nil {
+			internal.Logger.Printf(ctx, "sentinel: GetMasterAddrByName master=%q failed: %s",
+				c.opt.MasterName, err)
+			_ = sentinel.Close()
+			continue
+		}
+
+		// Push working sentinel to the top.
+		c.sentinelAddrs[0], c.sentinelAddrs[i] = c.sentinelAddrs[i], c.sentinelAddrs[0]
+		c.setSentinel(ctx, sentinel)
+
+		addr := net.JoinHostPort(masterAddr[0], masterAddr[1])
+		return addr, nil
+	}
+
+	return "", errors.New("redis: all sentinels are unreachable")
+}
+
+func (c *sentinelFailover) slaveAddrs(ctx context.Context) ([]string, error) {
+	c.mu.RLock()
+	sentinel := c.sentinel
+	c.mu.RUnlock()
+
+	if sentinel != nil {
+		addrs := c.getSlaveAddrs(ctx, sentinel)
+		if len(addrs) > 0 {
+			return addrs, nil
+		}
+	}
+
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if c.sentinel != nil {
+		addrs := c.getSlaveAddrs(ctx, c.sentinel)
+		if len(addrs) > 0 {
+			return addrs, nil
+		}
+		_ = c.closeSentinel()
+	}
+
+	for i, sentinelAddr := range c.sentinelAddrs {
+		sentinel := NewSentinelClient(c.opt.sentinelOptions(sentinelAddr))
+
+		slaves, err := sentinel.Slaves(ctx, c.opt.MasterName).Result()
+		if err != nil {
+			internal.Logger.Printf(ctx, "sentinel: Slaves master=%q failed: %s",
+				c.opt.MasterName, err)
+			_ = sentinel.Close()
+			continue
+		}
+
+		// Push working sentinel to the top.
+		c.sentinelAddrs[0], c.sentinelAddrs[i] = c.sentinelAddrs[i], c.sentinelAddrs[0]
+		c.setSentinel(ctx, sentinel)
+
+		addrs := parseSlaveAddrs(slaves)
+		return addrs, nil
+	}
+
+	return []string{}, errors.New("redis: all sentinels are unreachable")
+}
+
+func (c *sentinelFailover) getMasterAddr(ctx context.Context, sentinel *SentinelClient) string {
+	addr, err := sentinel.GetMasterAddrByName(ctx, c.opt.MasterName).Result()
+	if err != nil {
+		internal.Logger.Printf(ctx, "sentinel: GetMasterAddrByName name=%q failed: %s",
+			c.opt.MasterName, err)
+		return ""
+	}
+	return net.JoinHostPort(addr[0], addr[1])
+}
+
+func (c *sentinelFailover) getSlaveAddrs(ctx context.Context, sentinel *SentinelClient) []string {
+	addrs, err := sentinel.Slaves(ctx, c.opt.MasterName).Result()
+	if err != nil {
+		internal.Logger.Printf(ctx, "sentinel: Slaves name=%q failed: %s",
+			c.opt.MasterName, err)
+		return []string{}
+	}
+	return parseSlaveAddrs(addrs)
+}
+
+func parseSlaveAddrs(addrs []interface{}) []string {
+	nodes := make([]string, 0, len(addrs))
+
+	for _, node := range addrs {
+		ip := ""
+		port := ""
+		flags := []string{}
+		lastkey := ""
+		isDown := false
+
+		for _, key := range node.([]interface{}) {
+			switch lastkey {
+			case "ip":
+				ip = key.(string)
+			case "port":
+				port = key.(string)
+			case "flags":
+				flags = strings.Split(key.(string), ",")
+			}
+			lastkey = key.(string)
+		}
+
+		for _, flag := range flags {
+			switch flag {
+			case "s_down", "o_down", "disconnected":
+				isDown = true
+			}
+		}
+
+		if !isDown {
+			nodes = append(nodes, net.JoinHostPort(ip, port))
+		}
+	}
+
+	return nodes
+}
+
+func (c *sentinelFailover) trySwitchMaster(ctx context.Context, addr string) {
+	c.mu.RLock()
+	currentAddr := c._masterAddr
+	c.mu.RUnlock()
+
+	if addr == currentAddr {
+		return
+	}
+
+	c.mu.Lock()
+	defer c.mu.Unlock()
+
+	if addr == c._masterAddr {
+		return
+	}
+	c._masterAddr = addr
+
+	internal.Logger.Printf(ctx, "sentinel: new master=%q addr=%q",
+		c.opt.MasterName, addr)
+	if c.onFailover != nil {
+		c.onFailover(ctx, addr)
+	}
+}
+
+func (c *sentinelFailover) setSentinel(ctx context.Context, sentinel *SentinelClient) {
+	if c.sentinel != nil {
+		panic("not reached")
+	}
+	c.sentinel = sentinel
+	c.discoverSentinels(ctx)
+
+	c.pubsub = sentinel.Subscribe(ctx, "+switch-master", "+slave-reconf-done")
+	go c.listen(c.pubsub)
+}
+
+func (c *sentinelFailover) discoverSentinels(ctx context.Context) {
+	sentinels, err := c.sentinel.Sentinels(ctx, c.opt.MasterName).Result()
+	if err != nil {
+		internal.Logger.Printf(ctx, "sentinel: Sentinels master=%q failed: %s", c.opt.MasterName, err)
+		return
+	}
+	for _, sentinel := range sentinels {
+		vals := sentinel.([]interface{})
+		for i := 0; i < len(vals); i += 2 {
+			key := vals[i].(string)
+			if key == "name" {
+				sentinelAddr := vals[i+1].(string)
+				if !contains(c.sentinelAddrs, sentinelAddr) {
+					internal.Logger.Printf(ctx, "sentinel: discovered new sentinel=%q for master=%q",
+						sentinelAddr, c.opt.MasterName)
+					c.sentinelAddrs = append(c.sentinelAddrs, sentinelAddr)
+				}
+			}
+		}
+	}
+}
+
+func (c *sentinelFailover) listen(pubsub *PubSub) {
+	ctx := context.TODO()
+	if c.onUpdate != nil {
+		c.onUpdate(ctx)
+	}
+
+	ch := pubsub.Channel()
+	for msg := range ch {
+		if msg.Channel == "+switch-master" {
+			parts := strings.Split(msg.Payload, " ")
+			if parts[0] != c.opt.MasterName {
+				internal.Logger.Printf(pubsub.getContext(), "sentinel: ignore addr for master=%q", parts[0])
+				continue
+			}
+			addr := net.JoinHostPort(parts[3], parts[4])
+			c.trySwitchMaster(pubsub.getContext(), addr)
+		}
+
+		if c.onUpdate != nil {
+			c.onUpdate(ctx)
+		}
+	}
+}
+
+func contains(slice []string, str string) bool {
+	for _, s := range slice {
+		if s == str {
+			return true
+		}
+	}
+	return false
+}
+
+//------------------------------------------------------------------------------
+
+// NewFailoverClusterClient returns a client that supports routing read-only commands
+// to a slave node.
+func NewFailoverClusterClient(failoverOpt *FailoverOptions) *ClusterClient {
+	sentinelAddrs := make([]string, len(failoverOpt.SentinelAddrs))
+	copy(sentinelAddrs, failoverOpt.SentinelAddrs)
+
+	failover := &sentinelFailover{
+		opt:           failoverOpt,
+		sentinelAddrs: sentinelAddrs,
+	}
+
+	opt := failoverOpt.clusterOptions()
+	opt.ClusterSlots = func(ctx context.Context) ([]ClusterSlot, error) {
+		masterAddr, err := failover.MasterAddr(ctx)
+		if err != nil {
+			return nil, err
+		}
+
+		nodes := []ClusterNode{{
+			Addr: masterAddr,
+		}}
+
+		slaveAddrs, err := failover.slaveAddrs(ctx)
+		if err != nil {
+			return nil, err
+		}
+
+		for _, slaveAddr := range slaveAddrs {
+			nodes = append(nodes, ClusterNode{
+				Addr: slaveAddr,
+			})
+		}
+
+		slots := []ClusterSlot{
+			{
+				Start: 0,
+				End:   16383,
+				Nodes: nodes,
+			},
+		}
+		return slots, nil
+	}
+
+	c := NewClusterClient(opt)
+	failover.onUpdate = func(ctx context.Context) {
+		c.ReloadState(ctx)
+	}
+
+	return c
+}
diff --git a/vendor/github.com/go-redis/redis/v8/tx.go b/vendor/github.com/go-redis/redis/v8/tx.go
new file mode 100644
index 0000000..ad825c6
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/tx.go
@@ -0,0 +1,151 @@
+package redis
+
+import (
+	"context"
+
+	"github.com/go-redis/redis/v8/internal/pool"
+	"github.com/go-redis/redis/v8/internal/proto"
+)
+
+// TxFailedErr transaction redis failed.
+const TxFailedErr = proto.RedisError("redis: transaction failed")
+
+// Tx implements Redis transactions as described in
+// http://redis.io/topics/transactions. It's NOT safe for concurrent use
+// by multiple goroutines, because Exec resets list of watched keys.
+// If you don't need WATCH it is better to use Pipeline.
+type Tx struct {
+	baseClient
+	cmdable
+	statefulCmdable
+	hooks
+	ctx context.Context
+}
+
+func (c *Client) newTx(ctx context.Context) *Tx {
+	tx := Tx{
+		baseClient: baseClient{
+			opt:      c.opt,
+			connPool: pool.NewStickyConnPool(c.connPool),
+		},
+		hooks: c.hooks.clone(),
+		ctx:   ctx,
+	}
+	tx.init()
+	return &tx
+}
+
+func (c *Tx) init() {
+	c.cmdable = c.Process
+	c.statefulCmdable = c.Process
+}
+
+func (c *Tx) Context() context.Context {
+	return c.ctx
+}
+
+func (c *Tx) WithContext(ctx context.Context) *Tx {
+	if ctx == nil {
+		panic("nil context")
+	}
+	clone := *c
+	clone.init()
+	clone.hooks.lock()
+	clone.ctx = ctx
+	return &clone
+}
+
+func (c *Tx) Process(ctx context.Context, cmd Cmder) error {
+	return c.hooks.process(ctx, cmd, c.baseClient.process)
+}
+
+// Watch prepares a transaction and marks the keys to be watched
+// for conditional execution if there are any keys.
+//
+// The transaction is automatically closed when fn exits.
+func (c *Client) Watch(ctx context.Context, fn func(*Tx) error, keys ...string) error {
+	tx := c.newTx(ctx)
+	if len(keys) > 0 {
+		if err := tx.Watch(ctx, keys...).Err(); err != nil {
+			_ = tx.Close(ctx)
+			return err
+		}
+	}
+
+	err := fn(tx)
+	_ = tx.Close(ctx)
+	return err
+}
+
+// Close closes the transaction, releasing any open resources.
+func (c *Tx) Close(ctx context.Context) error {
+	_ = c.Unwatch(ctx).Err()
+	return c.baseClient.Close()
+}
+
+// Watch marks the keys to be watched for conditional execution
+// of a transaction.
+func (c *Tx) Watch(ctx context.Context, keys ...string) *StatusCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "watch"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Unwatch flushes all the previously watched keys for a transaction.
+func (c *Tx) Unwatch(ctx context.Context, keys ...string) *StatusCmd {
+	args := make([]interface{}, 1+len(keys))
+	args[0] = "unwatch"
+	for i, key := range keys {
+		args[1+i] = key
+	}
+	cmd := NewStatusCmd(ctx, args...)
+	_ = c.Process(ctx, cmd)
+	return cmd
+}
+
+// Pipeline creates a pipeline. Usually it is more convenient to use Pipelined.
+func (c *Tx) Pipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx: c.ctx,
+		exec: func(ctx context.Context, cmds []Cmder) error {
+			return c.hooks.processPipeline(ctx, cmds, c.baseClient.processPipeline)
+		},
+	}
+	pipe.init()
+	return &pipe
+}
+
+// Pipelined executes commands queued in the fn outside of the transaction.
+// Use TxPipelined if you need transactional behavior.
+func (c *Tx) Pipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.Pipeline().Pipelined(ctx, fn)
+}
+
+// TxPipelined executes commands queued in the fn in the transaction.
+//
+// When using WATCH, EXEC will execute commands only if the watched keys
+// were not modified, allowing for a check-and-set mechanism.
+//
+// Exec always returns list of commands. If transaction fails
+// TxFailedErr is returned. Otherwise Exec returns an error of the first
+// failed command or nil.
+func (c *Tx) TxPipelined(ctx context.Context, fn func(Pipeliner) error) ([]Cmder, error) {
+	return c.TxPipeline().Pipelined(ctx, fn)
+}
+
+// TxPipeline creates a pipeline. Usually it is more convenient to use TxPipelined.
+func (c *Tx) TxPipeline() Pipeliner {
+	pipe := Pipeline{
+		ctx: c.ctx,
+		exec: func(ctx context.Context, cmds []Cmder) error {
+			return c.hooks.processTxPipeline(ctx, cmds, c.baseClient.processTxPipeline)
+		},
+	}
+	pipe.init()
+	return &pipe
+}
diff --git a/vendor/github.com/go-redis/redis/v8/universal.go b/vendor/github.com/go-redis/redis/v8/universal.go
new file mode 100644
index 0000000..5f0e1e3
--- /dev/null
+++ b/vendor/github.com/go-redis/redis/v8/universal.go
@@ -0,0 +1,206 @@
+package redis
+
+import (
+	"context"
+	"crypto/tls"
+	"net"
+	"time"
+)
+
+// UniversalOptions information is required by UniversalClient to establish
+// connections.
+type UniversalOptions struct {
+	// Either a single address or a seed list of host:port addresses
+	// of cluster/sentinel nodes.
+	Addrs []string
+
+	// Database to be selected after connecting to the server.
+	// Only single-node and failover clients.
+	DB int
+
+	// Common options.
+
+	Dialer    func(ctx context.Context, network, addr string) (net.Conn, error)
+	OnConnect func(ctx context.Context, cn *Conn) error
+
+	Username         string
+	Password         string
+	SentinelPassword string
+
+	MaxRetries      int
+	MinRetryBackoff time.Duration
+	MaxRetryBackoff time.Duration
+
+	DialTimeout  time.Duration
+	ReadTimeout  time.Duration
+	WriteTimeout time.Duration
+
+	PoolSize           int
+	MinIdleConns       int
+	MaxConnAge         time.Duration
+	PoolTimeout        time.Duration
+	IdleTimeout        time.Duration
+	IdleCheckFrequency time.Duration
+
+	TLSConfig *tls.Config
+
+	// Only cluster clients.
+
+	MaxRedirects   int
+	ReadOnly       bool
+	RouteByLatency bool
+	RouteRandomly  bool
+
+	// The sentinel master name.
+	// Only failover clients.
+	MasterName string
+}
+
+// Cluster returns cluster options created from the universal options.
+func (o *UniversalOptions) Cluster() *ClusterOptions {
+	if len(o.Addrs) == 0 {
+		o.Addrs = []string{"127.0.0.1:6379"}
+	}
+
+	return &ClusterOptions{
+		Addrs:     o.Addrs,
+		Dialer:    o.Dialer,
+		OnConnect: o.OnConnect,
+
+		Username: o.Username,
+		Password: o.Password,
+
+		MaxRedirects:   o.MaxRedirects,
+		ReadOnly:       o.ReadOnly,
+		RouteByLatency: o.RouteByLatency,
+		RouteRandomly:  o.RouteRandomly,
+
+		MaxRetries:      o.MaxRetries,
+		MinRetryBackoff: o.MinRetryBackoff,
+		MaxRetryBackoff: o.MaxRetryBackoff,
+
+		DialTimeout:        o.DialTimeout,
+		ReadTimeout:        o.ReadTimeout,
+		WriteTimeout:       o.WriteTimeout,
+		PoolSize:           o.PoolSize,
+		MinIdleConns:       o.MinIdleConns,
+		MaxConnAge:         o.MaxConnAge,
+		PoolTimeout:        o.PoolTimeout,
+		IdleTimeout:        o.IdleTimeout,
+		IdleCheckFrequency: o.IdleCheckFrequency,
+
+		TLSConfig: o.TLSConfig,
+	}
+}
+
+// Failover returns failover options created from the universal options.
+func (o *UniversalOptions) Failover() *FailoverOptions {
+	if len(o.Addrs) == 0 {
+		o.Addrs = []string{"127.0.0.1:26379"}
+	}
+
+	return &FailoverOptions{
+		SentinelAddrs: o.Addrs,
+		MasterName:    o.MasterName,
+
+		Dialer:    o.Dialer,
+		OnConnect: o.OnConnect,
+
+		DB:               o.DB,
+		Username:         o.Username,
+		Password:         o.Password,
+		SentinelPassword: o.SentinelPassword,
+
+		MaxRetries:      o.MaxRetries,
+		MinRetryBackoff: o.MinRetryBackoff,
+		MaxRetryBackoff: o.MaxRetryBackoff,
+
+		DialTimeout:  o.DialTimeout,
+		ReadTimeout:  o.ReadTimeout,
+		WriteTimeout: o.WriteTimeout,
+
+		PoolSize:           o.PoolSize,
+		MinIdleConns:       o.MinIdleConns,
+		MaxConnAge:         o.MaxConnAge,
+		PoolTimeout:        o.PoolTimeout,
+		IdleTimeout:        o.IdleTimeout,
+		IdleCheckFrequency: o.IdleCheckFrequency,
+
+		TLSConfig: o.TLSConfig,
+	}
+}
+
+// Simple returns basic options created from the universal options.
+func (o *UniversalOptions) Simple() *Options {
+	addr := "127.0.0.1:6379"
+	if len(o.Addrs) > 0 {
+		addr = o.Addrs[0]
+	}
+
+	return &Options{
+		Addr:      addr,
+		Dialer:    o.Dialer,
+		OnConnect: o.OnConnect,
+
+		DB:       o.DB,
+		Username: o.Username,
+		Password: o.Password,
+
+		MaxRetries:      o.MaxRetries,
+		MinRetryBackoff: o.MinRetryBackoff,
+		MaxRetryBackoff: o.MaxRetryBackoff,
+
+		DialTimeout:  o.DialTimeout,
+		ReadTimeout:  o.ReadTimeout,
+		WriteTimeout: o.WriteTimeout,
+
+		PoolSize:           o.PoolSize,
+		MinIdleConns:       o.MinIdleConns,
+		MaxConnAge:         o.MaxConnAge,
+		PoolTimeout:        o.PoolTimeout,
+		IdleTimeout:        o.IdleTimeout,
+		IdleCheckFrequency: o.IdleCheckFrequency,
+
+		TLSConfig: o.TLSConfig,
+	}
+}
+
+// --------------------------------------------------------------------
+
+// UniversalClient is an abstract client which - based on the provided options -
+// can connect to either clusters, or sentinel-backed failover instances
+// or simple single-instance servers. This can be useful for testing
+// cluster-specific applications locally.
+type UniversalClient interface {
+	Cmdable
+	Context() context.Context
+	AddHook(Hook)
+	Watch(ctx context.Context, fn func(*Tx) error, keys ...string) error
+	Do(ctx context.Context, args ...interface{}) *Cmd
+	Process(ctx context.Context, cmd Cmder) error
+	Subscribe(ctx context.Context, channels ...string) *PubSub
+	PSubscribe(ctx context.Context, channels ...string) *PubSub
+	Close() error
+	PoolStats() *PoolStats
+}
+
+var (
+	_ UniversalClient = (*Client)(nil)
+	_ UniversalClient = (*ClusterClient)(nil)
+	_ UniversalClient = (*Ring)(nil)
+)
+
+// NewUniversalClient returns a new multi client. The type of client returned depends
+// on the following three conditions:
+//
+// 1. if a MasterName is passed a sentinel-backed FailoverClient will be returned
+// 2. if the number of Addrs is two or more, a ClusterClient will be returned
+// 3. otherwise, a single-node redis Client will be returned.
+func NewUniversalClient(opts *UniversalOptions) UniversalClient {
+	if opts.MasterName != "" {
+		return NewFailoverClient(opts.Failover())
+	} else if len(opts.Addrs) > 1 {
+		return NewClusterClient(opts.Cluster())
+	}
+	return NewClient(opts.Simple())
+}