EDGEPOD-35 Add Grafana dashboards for Aether
This patch introduces two new dashboards.
One is for monitoring overall resource usages and the other is for monitoring
Aether specific pods running in Edge.
Clusters are identified by "datacenter" external label.
Change-Id: I82fb2796fe54418a53817c70fcf20d1f61a4e059
diff --git a/nem-monitoring/Chart.yaml b/nem-monitoring/Chart.yaml
index 6fcd7b8..5fef10c 100644
--- a/nem-monitoring/Chart.yaml
+++ b/nem-monitoring/Chart.yaml
@@ -18,4 +18,4 @@
description: Time Series Storage and Dashboard for CORD
icon: https://guide.opencord.org/logos/cord.svg
-version: 1.1.0
+version: 1.2.0
diff --git a/nem-monitoring/grafana-dashboards/aether-clusters.json b/nem-monitoring/grafana-dashboards/aether-clusters.json
new file mode 100644
index 0000000..b558d2a
--- /dev/null
+++ b/nem-monitoring/grafana-dashboards/aether-clusters.json
@@ -0,0 +1,1709 @@
+{
+ "annotations": {
+ "list": [
+ {
+ "builtIn": 1,
+ "datasource": "-- Grafana --",
+ "enable": true,
+ "hide": true,
+ "iconColor": "rgba(0, 211, 255, 1)",
+ "name": "Annotations & Alerts",
+ "type": "dashboard"
+ }
+ ]
+ },
+ "description": "Monitors Kubernetes clusters using Prometheus.",
+ "editable": true,
+ "gnetId": 1621,
+ "graphTooltip": 0,
+ "iteration": 1574589072119,
+ "links": [],
+ "panels": [
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 0
+ },
+ "id": 45,
+ "panels": [],
+ "title": "Total usage in all data centers",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "cacheTimeout": null,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 0,
+ "y": 1
+ },
+ "id": 47,
+ "legend": {
+ "alignAsTable": true,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "links": [],
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pluginVersion": "6.4.2",
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "(1 - (avg(irate(node_cpu_seconds_total{mode=\"idle\",datacenter!=\"\"}[5m]))by (datacenter))) * 100",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "CPU usage (5m avg)",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "percent",
+ "label": null,
+ "logBase": 1,
+ "max": "100",
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "decimals": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 8,
+ "y": 1
+ },
+ "id": 49,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "(1 - sum(node_memory_MemAvailable_bytes{datacenter!=\"\"})by (datacenter)/ sum(node_memory_MemTotal_bytes{datacenter!=\"\"})by (datacenter)) * 100",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Memory usage",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "percent",
+ "label": null,
+ "logBase": 1,
+ "max": "100",
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 16,
+ "y": 1
+ },
+ "id": 51,
+ "legend": {
+ "alignAsTable": true,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum (rate (node_network_receive_bytes_total{datacenter!=\"\"}[5m]))by (datacenter)",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }} rx",
+ "refId": "A"
+ },
+ {
+ "expr": "- sum (rate (node_network_transmit_bytes_total{datacenter!=\"\"}[5m])) by (datacenter)",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }} tx",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Network I/O pressure",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 9
+ },
+ "id": 34,
+ "panels": [],
+ "title": "Total usage in $datacenter node $Node",
+ "type": "row"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": true,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": true,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 8,
+ "x": 0,
+ "y": 10
+ },
+ "height": "180px",
+ "id": 6,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "(1 - avg(irate(node_cpu_seconds_total{mode=\"idle\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"}[5m]))) * 100",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "65, 90",
+ "title": "CPU usage (5m avg)",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": true,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "editable": true,
+ "error": false,
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": true,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 8,
+ "x": 8,
+ "y": 10
+ },
+ "height": "180px",
+ "id": 4,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "(1 - sum(node_memory_MemAvailable_bytes{datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})/ sum(node_memory_MemTotal_bytes{datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})) * 100",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "65, 90",
+ "title": "Memory usage",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": true,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": true,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 8,
+ "x": 16,
+ "y": 10
+ },
+ "height": "180px",
+ "id": 7,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "(sum(node_filesystem_size_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})\n - sum(node_filesystem_free_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})) / sum(node_filesystem_size_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"}) * 100",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "",
+ "metric": "",
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "65, 90",
+ "title": "Filesystem usage",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 0,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 11,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": " cores",
+ "postfixFontSize": "30%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "(1 - (avg(irate(node_cpu_seconds_total{mode=\"idle\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"}[5m])))) * sum(machine_cpu_cores{datacenter=\"$datacenter\",kubernetes_io_hostname=~\"^$Node$\"})",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Used",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 4,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 12,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": " cores",
+ "postfixFontSize": "30%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum (machine_cpu_cores{datacenter=\"$datacenter\",kubernetes_io_hostname=~\"^$Node$\"})",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Total",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "bytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 8,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 9,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "20%",
+ "prefix": "",
+ "prefixFontSize": "20%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(node_memory_MemTotal_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node =~\"^$Node$\"}) - sum(node_memory_MemAvailable_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node =~\"^$Node$\"})",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Used",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "bytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 12,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 10,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum (machine_memory_bytes{datacenter=\"$datacenter\",kubernetes_io_hostname=~\"^$Node$\"})",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Total",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "bytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 16,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 13,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(node_filesystem_size_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})\n - sum(node_filesystem_free_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Used",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "format": "bytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 3,
+ "w": 4,
+ "x": 20,
+ "y": 15
+ },
+ "height": "1px",
+ "id": 14,
+ "interval": null,
+ "isNew": true,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(node_filesystem_size_bytes{device!~\"rootfs|HarddiskVolume.+\",datacenter=\"$datacenter\",kubernetes_node=~\"^$Node$\"})",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": "",
+ "title": "Total",
+ "type": "singlestat",
+ "valueFontSize": "50%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "0",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "fill": 1,
+ "fillGradient": 0,
+ "grid": {},
+ "gridPos": {
+ "h": 5,
+ "w": 24,
+ "x": 0,
+ "y": 18
+ },
+ "height": "200px",
+ "id": 32,
+ "isNew": true,
+ "legend": {
+ "alignAsTable": false,
+ "avg": true,
+ "current": true,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "sideWidth": 200,
+ "sort": "current",
+ "sortDesc": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 2,
+ "links": [],
+ "nullPointMode": "connected",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum (rate (node_network_receive_bytes_total{datacenter=\"$datacenter\"}[5m]))by (kubernetes_node)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "{{kubernetes_node}} Received",
+ "metric": "network",
+ "refId": "A",
+ "step": 10
+ },
+ {
+ "expr": "- sum (rate (node_network_transmit_bytes_total{datacenter=\"$datacenter\"}[5m])) by (kubernetes_node)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "{{kubernetes_node}} Sent",
+ "metric": "network",
+ "refId": "B",
+ "step": 10
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Network I/O pressure",
+ "tooltip": {
+ "msResolution": false,
+ "shared": true,
+ "sort": 0,
+ "value_type": "cumulative"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 23
+ },
+ "id": 35,
+ "panels": [],
+ "title": "Pods CPU usage in $datacenter namespace $namespace",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "Prometheus",
+ "decimals": 3,
+ "editable": true,
+ "error": false,
+ "fill": 1,
+ "fillGradient": 0,
+ "grid": {},
+ "gridPos": {
+ "h": 7,
+ "w": 24,
+ "x": 0,
+ "y": 24
+ },
+ "height": "",
+ "id": 17,
+ "isNew": true,
+ "legend": {
+ "alignAsTable": true,
+ "avg": true,
+ "current": true,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "sideWidth": null,
+ "sort": "avg",
+ "sortDesc": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 2,
+ "links": [],
+ "nullPointMode": "connected",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": true,
+ "targets": [
+ {
+ "expr": "sum (rate (container_cpu_usage_seconds_total{pod_name!=\"\", image!=\"\", container_name!=\"POD\", datacenter=\"$datacenter\", namespace=~\"^$namespace$\"}[5m])) by (pod_name)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "{{ pod_name }}",
+ "metric": "container_cpu",
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Pods CPU usage (5m avg)",
+ "tooltip": {
+ "msResolution": true,
+ "shared": true,
+ "sort": 2,
+ "value_type": "cumulative"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "percentunit",
+ "label": "cores",
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 31
+ },
+ "id": 39,
+ "panels": [],
+ "title": "Pods memory usage in $datacenter namespace $namespace",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "Prometheus",
+ "decimals": 2,
+ "editable": true,
+ "error": false,
+ "fill": 1,
+ "fillGradient": 0,
+ "grid": {},
+ "gridPos": {
+ "h": 7,
+ "w": 24,
+ "x": 0,
+ "y": 32
+ },
+ "id": 25,
+ "isNew": true,
+ "legend": {
+ "alignAsTable": true,
+ "avg": true,
+ "current": true,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "sideWidth": null,
+ "sort": "current",
+ "sortDesc": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 2,
+ "links": [],
+ "nullPointMode": "connected",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": true,
+ "targets": [
+ {
+ "expr": "sum(container_memory_working_set_bytes{container_name!=\"POD\", container_name!=\"\", namespace=~\"^$namespace$\", datacenter=\"$datacenter\"})by (pod_name)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "{{ pod_name }}",
+ "metric": "container_memory_usage:sort_desc",
+ "refId": "A",
+ "step": 10
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Pods memory usage",
+ "tooltip": {
+ "msResolution": false,
+ "shared": true,
+ "sort": 2,
+ "value_type": "cumulative"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "bytes",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 39
+ },
+ "id": 41,
+ "panels": [],
+ "title": "Pods network I/O in $datacenter namespace $namespace",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 7,
+ "w": 24,
+ "x": 0,
+ "y": 40
+ },
+ "id": 43,
+ "legend": {
+ "alignAsTable": true,
+ "avg": true,
+ "current": true,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum (rate (container_network_receive_bytes_total{pod_name!=\"\", container_name!=\"\", datacenter=\"$datacenter\",namespace=~\"^$namespace$\"}[5m])) by (pod_name)",
+ "legendFormat": "{{ pod_name }} Rcvd",
+ "refId": "A"
+ },
+ {
+ "expr": "- sum (rate (container_network_transmit_bytes_total{pod_name!=\"\", container_name!=\"\", datacenter=\"$datacenter\",namespace=~\"^$namespace$\"}[5m])) by (pod_name)",
+ "legendFormat": "{{ pod_name }} Sent",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "Pods network I/O (5m avg)",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "decimals": null,
+ "format": "Bps",
+ "label": "",
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ }
+ ],
+ "refresh": "1m",
+ "schemaVersion": 20,
+ "style": "dark",
+ "tags": [
+ "aether"
+ ],
+ "templating": {
+ "list": [
+ {
+ "allValue": null,
+ "datasource": "Prometheus",
+ "definition": "label_values(kube_pod_info, datacenter)",
+ "hide": 0,
+ "includeAll": false,
+ "label": "Datacenter",
+ "multi": false,
+ "name": "datacenter",
+ "options": [],
+ "query": "label_values(kube_pod_info, datacenter)",
+ "refresh": 2,
+ "regex": "",
+ "skipUrlSync": false,
+ "sort": 1,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": ".*",
+ "current": {
+ "tags": [],
+ "text": "All",
+ "value": "$__all"
+ },
+ "datasource": "Prometheus",
+ "definition": "label_values(node_load1{datacenter=\"$datacenter\"}, kubernetes_node)",
+ "hide": 0,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "Node",
+ "options": [],
+ "query": "label_values(node_load1{datacenter=\"$datacenter\"}, kubernetes_node)",
+ "refresh": 1,
+ "regex": "",
+ "skipUrlSync": false,
+ "sort": 0,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": ".*",
+ "current": {
+ "text": "All",
+ "value": "$__all"
+ },
+ "datasource": "Prometheus",
+ "definition": "label_values(kube_pod_info{datacenter=\"$datacenter\"}, namespace)",
+ "hide": 0,
+ "includeAll": true,
+ "label": "Namespace",
+ "multi": false,
+ "name": "namespace",
+ "options": [
+ {
+ "selected": true,
+ "text": "All",
+ "value": "$__all"
+ },
+ {
+ "selected": false,
+ "text": "default",
+ "value": "default"
+ },
+ {
+ "selected": false,
+ "text": "kube-system",
+ "value": "kube-system"
+ },
+ {
+ "selected": false,
+ "text": "omec",
+ "value": "omec"
+ }
+ ],
+ "query": "label_values(kube_pod_info{datacenter=\"$datacenter\"}, namespace)",
+ "refresh": 0,
+ "regex": "",
+ "skipUrlSync": false,
+ "sort": 0,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ }
+ ]
+ },
+ "time": {
+ "from": "now-30m",
+ "to": "now"
+ },
+ "timepicker": {
+ "refresh_intervals": [
+ "5s",
+ "10s",
+ "30s",
+ "1m",
+ "5m",
+ "15m",
+ "30m",
+ "1h",
+ "2h",
+ "1d"
+ ],
+ "time_options": [
+ "5m",
+ "15m",
+ "1h",
+ "6h",
+ "12h",
+ "24h",
+ "2d",
+ "7d",
+ "30d"
+ ]
+ },
+ "timezone": "browser",
+ "title": "Aether Clusters",
+ "uid": "icjpCppik",
+ "version": 6
+}
\ No newline at end of file
diff --git a/nem-monitoring/grafana-dashboards/aether-edge-pods.json b/nem-monitoring/grafana-dashboards/aether-edge-pods.json
new file mode 100644
index 0000000..4158679
--- /dev/null
+++ b/nem-monitoring/grafana-dashboards/aether-edge-pods.json
@@ -0,0 +1,1086 @@
+{
+ "annotations": {
+ "list": [
+ {
+ "builtIn": 1,
+ "datasource": "-- Grafana --",
+ "enable": true,
+ "hide": true,
+ "iconColor": "rgba(0, 211, 255, 1)",
+ "name": "Annotations & Alerts",
+ "type": "dashboard"
+ }
+ ]
+ },
+ "editable": true,
+ "gnetId": null,
+ "graphTooltip": 0,
+ "id": 3,
+ "links": [],
+ "panels": [
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 0
+ },
+ "id": 4,
+ "panels": [],
+ "repeat": "pod",
+ "scopedVars": {
+ "pod": {
+ "selected": true,
+ "text": "spgwu",
+ "value": "spgwu"
+ }
+ },
+ "title": "$pod",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 0,
+ "y": 1
+ },
+ "id": 16,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "scopedVars": {
+ "pod": {
+ "selected": true,
+ "text": "spgwu",
+ "value": "spgwu"
+ }
+ },
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": true,
+ "targets": [
+ {
+ "expr": "sum (rate (container_cpu_usage_seconds_total{image!=\"\", container_name!=\"POD\", pod_name=\"$pod-0\"}[5m])) by (datacenter)",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "$pod CPU usage in all Edges",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "decimals": 2,
+ "format": "percentunit",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 8,
+ "y": 1
+ },
+ "id": 17,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "scopedVars": {
+ "pod": {
+ "selected": true,
+ "text": "spgwu",
+ "value": "spgwu"
+ }
+ },
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(container_memory_working_set_bytes{container_name!=\"POD\", container_name!=\"\", pod_name=\"$pod-0\"}) by (datacenter)",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "$pod memory usage in all Edges",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "decimals": 2,
+ "format": "bytes",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 16,
+ "y": 1
+ },
+ "id": 18,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 2,
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 2,
+ "points": false,
+ "renderer": "flot",
+ "scopedVars": {
+ "pod": {
+ "selected": true,
+ "text": "spgwu",
+ "value": "spgwu"
+ }
+ },
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(rate(container_network_receive_bytes_total{pod_name=\"$pod-0\",container_name!=\"\"}[5m]))by (datacenter)",
+ "interval": "10s",
+ "legendFormat": "{{ datacenter }} rx",
+ "refId": "A"
+ },
+ {
+ "expr": "- sum(rate(container_network_transmit_bytes_total{pod_name=\"$pod-0\",container_name!=\"\"}[5m]))by (datacenter)",
+ "legendFormat": "{{ datacenter }} tx",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "$pod network I/O in all Edges",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "collapsed": false,
+ "datasource": null,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 27
+ },
+ "id": 25,
+ "panels": [],
+ "repeat": "datacenter",
+ "title": "$pod usage in $datacenter",
+ "type": "row"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": true,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": null,
+ "description": "",
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": true,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 6,
+ "x": 0,
+ "y": 28
+ },
+ "id": 10,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "maxPerRow": 12,
+ "nullPointMode": "null",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": true
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "100 * sum(irate(container_cpu_usage_seconds_total{pod_name=\"$pod-0\", image!=\"\", container_name!=\"POD\", datacenter=\"$datacenter\"}[3m])) by (pod_name) / sum(container_spec_cpu_quota{image!=\"\", container_name!=\"POD\", datacenter=\"$datacenter\"} / container_spec_cpu_period{image!=\"\", container_name!=\"POD\", datacenter=\"$datacenter\"}) by (pod_name)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "",
+ "refId": "A"
+ }
+ ],
+ "thresholds": "85,95",
+ "title": "$pod CPU usage in $datacenter",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": true,
+ "colors": [
+ "rgba(50, 172, 45, 0.97)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(245, 54, 54, 0.9)"
+ ],
+ "datasource": "Prometheus",
+ "decimals": null,
+ "description": "",
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": true,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 6,
+ "x": 6,
+ "y": 28
+ },
+ "id": 12,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "maxPerRow": 12,
+ "nullPointMode": "null",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": true
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "100 * (sum(container_memory_usage_bytes{pod_name=\"$pod-0\", container_name!=\"POD\", datacenter=\"$datacenter\"}) / sum(container_spec_memory_limit_bytes{pod_name=\"$pod-0\", container_name!=\"POD\", datacenter=\"$datacenter\"}))",
+ "format": "time_series",
+ "hide": false,
+ "instant": false,
+ "interval": "10s",
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "85,95",
+ "title": "$pod memory usage in $datacenter",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 2,
+ "fillGradient": 0,
+ "gridPos": {
+ "h": 9,
+ "w": 12,
+ "x": 12,
+ "y": 28
+ },
+ "id": 2,
+ "legend": {
+ "alignAsTable": true,
+ "avg": true,
+ "current": true,
+ "max": false,
+ "min": false,
+ "rightSide": true,
+ "show": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 2,
+ "links": [],
+ "nullPointMode": "null",
+ "options": {
+ "dataLinks": []
+ },
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "repeat": null,
+ "repeatDirection": "v",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(rate(container_network_receive_bytes_total{pod_name=\"$pod-0\", datacenter=\"$datacenter\"}[3m]))by (interface)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "rx: {{ interface }}",
+ "refId": "B"
+ },
+ {
+ "expr": "-sum(rate(container_network_transmit_bytes_total{pod_name=\"$pod-0\", datacenter=\"$datacenter\"}[3m]))by (interface)",
+ "format": "time_series",
+ "interval": "10s",
+ "intervalFactor": 1,
+ "legendFormat": "tx: {{ interface }}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeRegions": [],
+ "timeShift": null,
+ "title": "$pod network I/O in $datacenter",
+ "tooltip": {
+ "shared": true,
+ "sort": 1,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "Bps",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ],
+ "yaxis": {
+ "align": false,
+ "alignLevel": null
+ }
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 4,
+ "w": 3,
+ "x": 0,
+ "y": 33
+ },
+ "id": 20,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false,
+ "ymax": null,
+ "ymin": null
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(kube_pod_container_resource_limits_cpu_cores{datacenter=\"$datacenter\", pod=\"$pod-0\"})",
+ "interval": "10s",
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "CPU limits",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "avg"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 4,
+ "w": 3,
+ "x": 3,
+ "y": 33
+ },
+ "id": 21,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false,
+ "ymax": null,
+ "ymin": null
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum (rate (container_cpu_usage_seconds_total{image!=\"\", container_name!=\"POD\", datacenter=\"$datacenter\", pod_name=\"$pod-0\"}[5m]))",
+ "interval": "10s",
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "CPU used (5m avg)",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "avg"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "decimals": 2,
+ "format": "decbytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 4,
+ "w": 3,
+ "x": 6,
+ "y": 33
+ },
+ "id": 22,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false,
+ "ymax": null,
+ "ymin": null
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(kube_pod_container_resource_limits_memory_bytes{datacenter=\"$datacenter\", pod=\"$pod-0\"})",
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Memory limits",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "avg"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "format": "decbytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 4,
+ "w": 3,
+ "x": 9,
+ "y": 33
+ },
+ "id": 23,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "options": {},
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "repeat": null,
+ "repeatDirection": "v",
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false,
+ "ymax": null,
+ "ymin": null
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum(container_memory_working_set_bytes{container_name!=\"POD\", container_name!=\"\", datacenter=\"$datacenter\", pod_name=\"$pod-0\"})",
+ "interval": "10s",
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Memory used",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "avg"
+ }
+ ],
+ "refresh": "5s",
+ "schemaVersion": 20,
+ "style": "dark",
+ "tags": [
+ "aether"
+ ],
+ "templating": {
+ "list": [
+ {
+ "allValue": ".*",
+ "current": {
+ "text": "spgwu",
+ "value": [
+ "spgwu"
+ ]
+ },
+ "hide": 0,
+ "includeAll": false,
+ "label": "Pod",
+ "multi": false,
+ "name": "pod",
+ "options": [
+ {
+ "selected": true,
+ "text": "spgwu",
+ "value": "spgwu"
+ },
+ {
+ "selected": false,
+ "text": "accelleran-cbrs-ric",
+ "value": "accelleran-cbrs-ric"
+ },
+ {
+ "selected": false,
+ "text": "accelleran-cbrs-nats",
+ "value": "accelleran-cbrs-nats"
+ },
+ {
+ "selected": false,
+ "text": "accelleran-cbrs-redis",
+ "value": "accelleran-cbrs-redis"
+ },
+ {
+ "selected": false,
+ "text": "accelleran-cbrs-ricweb",
+ "value": "accelleran-cbrs-ricweb"
+ }
+ ],
+ "query": "spgwu,accelleran-cbrs-ric,accelleran-cbrs-nats,accelleran-cbrs-redis,accelleran-cbrs-ricweb",
+ "skipUrlSync": false,
+ "type": "custom"
+ },
+ {
+ "allValue": ".*",
+ "current": {
+ "text": "All",
+ "value": [
+ "$__all"
+ ]
+ },
+ "datasource": "Prometheus",
+ "definition": "label_values(kube_pod_info, datacenter)",
+ "hide": 0,
+ "includeAll": true,
+ "label": "Datacenter",
+ "multi": true,
+ "name": "datacenter",
+ "options": [],
+ "query": "label_values(kube_pod_info, datacenter)",
+ "refresh": 2,
+ "regex": "",
+ "skipUrlSync": false,
+ "sort": 1,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ }
+ ]
+ },
+ "time": {
+ "from": "now-5m",
+ "to": "now"
+ },
+ "timepicker": {
+ "hidden": false,
+ "refresh_intervals": [
+ "5s",
+ "10s",
+ "30s",
+ "1m",
+ "5m",
+ "15m",
+ "30m",
+ "1h",
+ "2h",
+ "1d"
+ ],
+ "time_options": [
+ "5m",
+ "15m",
+ "1h",
+ "6h",
+ "12h",
+ "24h",
+ "2d",
+ "7d",
+ "30d"
+ ]
+ },
+ "timezone": "",
+ "title": "Aether Edge Pods",
+ "uid": "gKU2e27Za",
+ "version": 3
+}
\ No newline at end of file
diff --git a/nem-monitoring/templates/grafana-dashboard-configmap.yaml b/nem-monitoring/templates/grafana-dashboard-configmap.yaml
index c217d20..47382ed 100644
--- a/nem-monitoring/templates/grafana-dashboard-configmap.yaml
+++ b/nem-monitoring/templates/grafana-dashboard-configmap.yaml
@@ -48,3 +48,9 @@
omec.json: |
{{ .Files.Get "grafana-dashboards/omec.json" | indent 4 }}
{{- end }}
+{{- if .Values.dashboards.aether }}
+ aether-clusters.json: |
+{{ .Files.Get "grafana-dashboards/aether-clusters.json" | indent 4 }}
+ aether-edge-pods.json: |
+{{ .Files.Get "grafana-dashboards/aether-edge-pods.json" | indent 4 }}
+{{- end }}
diff --git a/nem-monitoring/values.yaml b/nem-monitoring/values.yaml
index 6d9220e..6d123bf 100644
--- a/nem-monitoring/values.yaml
+++ b/nem-monitoring/values.yaml
@@ -96,3 +96,4 @@
aaa: true
voltha: true
omec: false
+ aether: false