53234ab0e5
Since these services are replicated (two per service), we should check for the max, not min, as one of the replicas could be down, but still the API is Up. Change-Id: I6dad9b74d65415f5f4f91e00c732792010487619
737 lines
19 KiB
JSON
737 lines
19 KiB
JSON
{% raw %}
|
|
{
|
|
"annotations": {
|
|
"list": []
|
|
},
|
|
"description": "Monitors Alertmanager using Prometheus.",
|
|
"editable": true,
|
|
"gnetId": 315,
|
|
"graphTooltip": 1,
|
|
"hideControls": false,
|
|
"id": null,
|
|
"links": [],
|
|
"refresh": "1m",
|
|
"rows": [
|
|
{
|
|
"collapse": false,
|
|
"height": 250,
|
|
"panels": [
|
|
{
|
|
"cacheTimeout": null,
|
|
"colorBackground": false,
|
|
"colorValue": true,
|
|
"colors": [
|
|
"rgba(245, 54, 54, 0.9)",
|
|
"rgba(237, 129, 40, 0.89)",
|
|
"rgba(50, 172, 45, 0.97)"
|
|
],
|
|
"datasource": "prometheus",
|
|
"format": "none",
|
|
"gauge": {
|
|
"maxValue": 1,
|
|
"minValue": 0,
|
|
"show": false,
|
|
"thresholdLabels": false,
|
|
"thresholdMarkers": true
|
|
},
|
|
"id": 1,
|
|
"interval": null,
|
|
"links": [],
|
|
"mappingType": 1,
|
|
"mappingTypes": [
|
|
{
|
|
"name": "value to text",
|
|
"value": 1
|
|
},
|
|
{
|
|
"name": "range to text",
|
|
"value": 2
|
|
}
|
|
],
|
|
"maxDataPoints": 100,
|
|
"nullPointMode": "connected",
|
|
"nullText": null,
|
|
"postfix": "",
|
|
"postfixFontSize": "50%",
|
|
"prefix": "",
|
|
"prefixFontSize": "50%",
|
|
"rangeMaps": [
|
|
{
|
|
"from": "null",
|
|
"text": "N/A",
|
|
"to": "null"
|
|
}
|
|
],
|
|
"span": 2,
|
|
"sparkline": {
|
|
"fillColor": "rgba(31, 118, 189, 0.18)",
|
|
"full": false,
|
|
"lineColor": "rgb(31, 120, 193)",
|
|
"show": true
|
|
},
|
|
"tableColumn": "",
|
|
"targets": [
|
|
{
|
|
"expr": "max(up{job=\"alertmanager\"})",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "",
|
|
"refId": "A",
|
|
"step": 60
|
|
}
|
|
],
|
|
"thresholds": "1,1",
|
|
"title": "Status",
|
|
"type": "singlestat",
|
|
"valueFontSize": "80%",
|
|
"valueMaps": [
|
|
{
|
|
"op": "=",
|
|
"text": "UNKNOWN",
|
|
"value": "null"
|
|
},
|
|
{
|
|
"op": "=",
|
|
"text": "OK",
|
|
"value": "1"
|
|
},
|
|
{
|
|
"op": "=",
|
|
"text": "DOWN",
|
|
"value": "0"
|
|
}
|
|
],
|
|
"valueName": "current"
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"decimals": 0,
|
|
"fill": 1,
|
|
"id": 2,
|
|
"legend": {
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"span": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "rate(alertmanager_alerts_received_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[10m])",
|
|
"format": "time_series",
|
|
"hide": false,
|
|
"intervalFactor": 2,
|
|
"legendFormat": "10m rate {{ status }}",
|
|
"refId": "A",
|
|
"step": 4
|
|
},
|
|
{
|
|
"expr": "rate(alertmanager_alerts_received_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[1h])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "1h rate {{ status }}",
|
|
"refId": "B",
|
|
"step": 4
|
|
},
|
|
{
|
|
"expr": "rate(alertmanager_alerts_received_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[6h])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "6h rate {{ status }}",
|
|
"refId": "C",
|
|
"step": 4
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "Received alerts",
|
|
"tooltip": {
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": "0",
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"repeat": null,
|
|
"repeatIteration": null,
|
|
"repeatRowId": null,
|
|
"showTitle": true,
|
|
"title": "Service Status",
|
|
"titleSize": "h6"
|
|
},
|
|
{
|
|
"collapse": false,
|
|
"height": 250,
|
|
"panels": [
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"fill": 1,
|
|
"id": 11,
|
|
"legend": {
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"span": 6,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "rate(alertmanager_notifications_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[5m])\n",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ integration }}",
|
|
"refId": "A",
|
|
"step": 10
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "Successful notifications rate",
|
|
"tooltip": {
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"fill": 1,
|
|
"id": 10,
|
|
"legend": {
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"span": 6,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "rate(alertmanager_notifications_failed_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[5m])\n",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ integration }}",
|
|
"refId": "A",
|
|
"step": 10
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "Failed notifications rate",
|
|
"tooltip": {
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"repeat": null,
|
|
"repeatIteration": null,
|
|
"repeatRowId": null,
|
|
"showTitle": true,
|
|
"title": "Alerts",
|
|
"titleSize": "h6"
|
|
},
|
|
{
|
|
"collapse": false,
|
|
"height": 250,
|
|
"panels": [
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"decimals": null,
|
|
"editable": true,
|
|
"error": false,
|
|
"fill": 1,
|
|
"id": 6,
|
|
"legend": {
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [
|
|
{
|
|
"alias": "/-/",
|
|
"fill": 0
|
|
}
|
|
],
|
|
"spaceLength": 10,
|
|
"span": 4,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "process_resident_memory_bytes{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "RSS",
|
|
"metric": "process_resident_memory_bytes",
|
|
"refId": "B",
|
|
"step": 20
|
|
},
|
|
{
|
|
"expr": "max_over_time(go_memstats_alloc_bytes{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[5m])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "Allocated bytes - 5m max",
|
|
"metric": "go_memstats_alloc_bytes",
|
|
"refId": "D",
|
|
"step": 20
|
|
},
|
|
{
|
|
"expr": "go_memstats_alloc_bytes{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "Allocated bytes",
|
|
"metric": "go_memstats_alloc_bytes",
|
|
"refId": "A",
|
|
"step": 20
|
|
},
|
|
{
|
|
"expr": "min_over_time(go_memstats_alloc_bytes{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[5m])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "Allocated bytes - 5m min",
|
|
"metric": "go_memstats_alloc_bytes",
|
|
"refId": "C",
|
|
"step": 20
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "Memory",
|
|
"tooltip": {
|
|
"msResolution": false,
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "bytes",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": "0",
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"editable": true,
|
|
"error": false,
|
|
"fill": 1,
|
|
"id": 7,
|
|
"legend": {
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"span": 4,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "rate(go_memstats_alloc_bytes_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[1m])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "Allocated Bytes/s",
|
|
"metric": "go_memstats_alloc_bytes",
|
|
"refId": "A",
|
|
"step": 20
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "Allocations",
|
|
"tooltip": {
|
|
"msResolution": false,
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "bytes",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": "0",
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": "prometheus",
|
|
"decimals": 2,
|
|
"editable": true,
|
|
"error": false,
|
|
"fill": 1,
|
|
"id": 9,
|
|
"legend": {
|
|
"alignAsTable": false,
|
|
"avg": false,
|
|
"current": false,
|
|
"hideEmpty": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": false,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"span": 4,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"expr": "irate(process_cpu_seconds_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[1m])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "Irate",
|
|
"metric": "",
|
|
"refId": "A",
|
|
"step": 20
|
|
},
|
|
{
|
|
"expr": "rate(process_cpu_seconds_total{job=\"alertmanager\",instance=~\"$Alertmanager:[1-9][0-9]*\"}[5m])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "5m rate",
|
|
"metric": "",
|
|
"refId": "B",
|
|
"step": 20
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeFrom": null,
|
|
"timeShift": null,
|
|
"title": "CPU",
|
|
"tooltip": {
|
|
"msResolution": false,
|
|
"shared": true,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"buckets": null,
|
|
"mode": "time",
|
|
"name": null,
|
|
"show": true,
|
|
"values": [
|
|
"avg"
|
|
]
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "none",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": "0",
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"label": null,
|
|
"logBase": 1,
|
|
"max": null,
|
|
"min": null,
|
|
"show": true
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"repeat": null,
|
|
"repeatIteration": null,
|
|
"repeatRowId": null,
|
|
"showTitle": true,
|
|
"title": "Memory & CPU",
|
|
"titleSize": "h6"
|
|
}
|
|
],
|
|
"schemaVersion": 14,
|
|
"style": "dark",
|
|
"tags": [
|
|
"prometheus"
|
|
],
|
|
"templating": {
|
|
"list": [
|
|
{
|
|
"allValue": null,
|
|
"current": {},
|
|
"datasource": "prometheus",
|
|
"hide": 0,
|
|
"includeAll": false,
|
|
"label": "Alertmanager",
|
|
"multi": false,
|
|
"name": "Alertmanager",
|
|
"options": [],
|
|
"query": "query_result(up{job=\"alertmanager\"} == 1)",
|
|
"refresh": 1,
|
|
"refresh_on_load": true,
|
|
"regex": ".*instance=\"([^:]+):[1-9][0-9]*\".*",
|
|
"sort": 1,
|
|
"tagValuesQuery": null,
|
|
"tags": [],
|
|
"tagsQuery": null,
|
|
"type": "query",
|
|
"useTags": false
|
|
}
|
|
]
|
|
},
|
|
"time": {
|
|
"from": "now-1h",
|
|
"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": "Alertmanager",
|
|
"version": 1
|
|
}
|
|
{%- endraw %}
|