mirror of
https://github.com/storytold/terraform-aws-observability-accelerator.git
synced 2026-10-09 00:09:43 +00:00
Rodrigue Koffi
1586 lines
39 KiB
JSON
1586 lines
39 KiB
JSON
{
|
|
"annotations": {
|
|
"list": [
|
|
{
|
|
"builtIn": 1,
|
|
"datasource": {
|
|
"type": "grafana",
|
|
"uid": "-- Grafana --"
|
|
},
|
|
"enable": true,
|
|
"hide": true,
|
|
"iconColor": "rgba(0, 211, 255, 1)",
|
|
"name": "Annotations & Alerts",
|
|
"target": {
|
|
"limit": 100,
|
|
"matchAny": false,
|
|
"tags": [],
|
|
"type": "dashboard"
|
|
},
|
|
"type": "dashboard"
|
|
}
|
|
]
|
|
},
|
|
"editable": false,
|
|
"fiscalYearStartMonth": 0,
|
|
"graphTooltip": 0,
|
|
"id": 23,
|
|
"iteration": 1656211483985,
|
|
"links": [],
|
|
"liveNow": false,
|
|
"panels": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"description": "The SLO (service level objective) and other metrics displayed on this dashboard are for informational purposes only.",
|
|
"gridPos": {
|
|
"h": 2,
|
|
"w": 24,
|
|
"x": 0,
|
|
"y": 0
|
|
},
|
|
"id": 2,
|
|
"options": {
|
|
"content": "The SLO (service level objective) and other metrics displayed on this dashboard are for informational purposes only.",
|
|
"mode": "markdown"
|
|
},
|
|
"pluginVersion": "9.0.1",
|
|
"span": 12,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"title": "Notice",
|
|
"type": "text"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many percent of requests (both read and write) in 30 days have been answered successfully and fast enough?",
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"decimals": 3,
|
|
"mappings": [
|
|
{
|
|
"options": {
|
|
"match": "null",
|
|
"result": {
|
|
"text": "N/A"
|
|
}
|
|
},
|
|
"type": "special"
|
|
}
|
|
],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 0,
|
|
"y": 2
|
|
},
|
|
"id": 3,
|
|
"interval": "1m",
|
|
"links": [],
|
|
"maxDataPoints": 100,
|
|
"options": {
|
|
"colorMode": "none",
|
|
"graphMode": "none",
|
|
"justifyMode": "auto",
|
|
"orientation": "horizontal",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"mean"
|
|
],
|
|
"fields": "",
|
|
"values": false
|
|
},
|
|
"textMode": "auto"
|
|
},
|
|
"pluginVersion": "9.0.1",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "apiserver_request:availability30d{verb=\"all\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"title": "Availability (30d) > 99.000%",
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"decimals": 3,
|
|
"description": "How much error budget is left looking at our 0.990% availability guarantees?",
|
|
"fill": 10,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 16,
|
|
"x": 8,
|
|
"y": 2
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 4,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "100 * (apiserver_request:availability30d{verb=\"all\", cluster=\"$cluster\"} - 0.990000)",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "errorbudget",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "ErrorBudget (30d) > 99.000%",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"decimals": 3,
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"decimals": 3,
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many percent of read requests (LIST,GET) in 30 days have been answered successfully and fast enough?",
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"decimals": 3,
|
|
"mappings": [
|
|
{
|
|
"options": {
|
|
"match": "null",
|
|
"result": {
|
|
"text": "N/A"
|
|
}
|
|
},
|
|
"type": "special"
|
|
}
|
|
],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 0,
|
|
"y": 9
|
|
},
|
|
"id": 5,
|
|
"interval": "1m",
|
|
"links": [],
|
|
"maxDataPoints": 100,
|
|
"options": {
|
|
"colorMode": "none",
|
|
"graphMode": "none",
|
|
"justifyMode": "auto",
|
|
"orientation": "horizontal",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"mean"
|
|
],
|
|
"fields": "",
|
|
"values": false
|
|
},
|
|
"textMode": "auto"
|
|
},
|
|
"pluginVersion": "9.0.1",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "apiserver_request:availability30d{verb=\"read\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"title": "Read Availability (30d)",
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many read requests (LIST,GET) per second do the apiservers get by code?",
|
|
"fill": 10,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 6,
|
|
"y": 9
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 6,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [
|
|
{
|
|
"alias": "/2../i",
|
|
"color": "#56A64B"
|
|
},
|
|
{
|
|
"alias": "/3../i",
|
|
"color": "#F2CC0C"
|
|
},
|
|
{
|
|
"alias": "/4../i",
|
|
"color": "#3274D9"
|
|
},
|
|
{
|
|
"alias": "/5../i",
|
|
"color": "#E02F44"
|
|
}
|
|
],
|
|
"spaceLength": 10,
|
|
"stack": true,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum by (code) (code_resource:apiserver_request_total:rate5m{verb=\"read\", cluster=\"$cluster\"})",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ code }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Read SLI - Requests",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "reqps",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "reqps",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many percent of read requests (LIST,GET) per second are returned with errors (5xx)?",
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 12,
|
|
"y": 9
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 7,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum by (resource) (code_resource:apiserver_request_total:rate5m{verb=\"read\",code=~\"5..\", cluster=\"$cluster\"}) / sum by (resource) (code_resource:apiserver_request_total:rate5m{verb=\"read\", cluster=\"$cluster\"})",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ resource }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Read SLI - Errors",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many seconds is the 99th percentile for reading (LIST|GET) a given resource?",
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 18,
|
|
"y": 9
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 8,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "cluster_quantile:apiserver_request_slo_duration_seconds:histogram_quantile{verb=\"read\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ resource }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Read SLI - Duration",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many percent of write requests (POST|PUT|PATCH|DELETE) in 30 days have been answered successfully and fast enough?",
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"decimals": 3,
|
|
"mappings": [
|
|
{
|
|
"options": {
|
|
"match": "null",
|
|
"result": {
|
|
"text": "N/A"
|
|
}
|
|
},
|
|
"type": "special"
|
|
}
|
|
],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 0,
|
|
"y": 16
|
|
},
|
|
"id": 9,
|
|
"interval": "1m",
|
|
"links": [],
|
|
"maxDataPoints": 100,
|
|
"options": {
|
|
"colorMode": "none",
|
|
"graphMode": "none",
|
|
"justifyMode": "auto",
|
|
"orientation": "horizontal",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"mean"
|
|
],
|
|
"fields": "",
|
|
"values": false
|
|
},
|
|
"textMode": "auto"
|
|
},
|
|
"pluginVersion": "9.0.1",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "apiserver_request:availability30d{verb=\"write\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"title": "Write Availability (30d)",
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many write requests (POST|PUT|PATCH|DELETE) per second do the apiservers get by code?",
|
|
"fill": 10,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 6,
|
|
"y": 16
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 10,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [
|
|
{
|
|
"alias": "/2../i",
|
|
"color": "#56A64B"
|
|
},
|
|
{
|
|
"alias": "/3../i",
|
|
"color": "#F2CC0C"
|
|
},
|
|
{
|
|
"alias": "/4../i",
|
|
"color": "#3274D9"
|
|
},
|
|
{
|
|
"alias": "/5../i",
|
|
"color": "#E02F44"
|
|
}
|
|
],
|
|
"spaceLength": 10,
|
|
"stack": true,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum by (code) (code_resource:apiserver_request_total:rate5m{verb=\"write\", cluster=\"$cluster\"})",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ code }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Write SLI - Requests",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "reqps",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "reqps",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many percent of write requests (POST|PUT|PATCH|DELETE) per second are returned with errors (5xx)?",
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 12,
|
|
"y": 16
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 11,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum by (resource) (code_resource:apiserver_request_total:rate5m{verb=\"write\",code=~\"5..\", cluster=\"$cluster\"}) / sum by (resource) (code_resource:apiserver_request_total:rate5m{verb=\"write\", cluster=\"$cluster\"})",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ resource }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Write SLI - Errors",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "percentunit",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"description": "How many seconds is the 99th percentile for writing (POST|PUT|PATCH|DELETE) a given resource?",
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 6,
|
|
"x": 18,
|
|
"y": 16
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 12,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "cluster_quantile:apiserver_request_slo_duration_seconds:histogram_quantile{verb=\"write\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{ resource }}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Write SLI - Duration",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 12,
|
|
"x": 0,
|
|
"y": 23
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 13,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": false,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum(rate(workqueue_adds_total{job=\"apiserver\", instance=~\"$instance\", cluster=\"$cluster\"}[$__rate_interval])) by (instance, name)",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}} {{name}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Work Queue Add Rate",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "ops",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "ops",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 12,
|
|
"x": 12,
|
|
"y": 23
|
|
},
|
|
"hiddenSeries": false,
|
|
"id": 14,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": false,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"options": {
|
|
"alertThreshold": true
|
|
},
|
|
"percentage": false,
|
|
"pluginVersion": "9.0.1",
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "sum(rate(workqueue_depth{job=\"apiserver\", instance=~\"$instance\", cluster=\"$cluster\"}[$__rate_interval])) by (instance, name)",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}} {{name}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"timeRegions": [],
|
|
"title": "Work Queue Depth",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
}
|
|
],
|
|
"yaxis": {
|
|
"align": false
|
|
}
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 24,
|
|
"x": 0,
|
|
"y": 30
|
|
},
|
|
"id": 15,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": true,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": true
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "histogram_quantile(0.99, sum(rate(workqueue_queue_duration_seconds_bucket{job=\"apiserver\", instance=~\"$instance\", cluster=\"$cluster\"}[$__rate_interval])) by (instance, name, le))",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}} {{name}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"title": "Work Queue Latency",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "s",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 0,
|
|
"y": 37
|
|
},
|
|
"id": 16,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "process_resident_memory_bytes{job=\"apiserver\",instance=~\"$instance\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"title": "Memory",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "bytes",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "bytes",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 8,
|
|
"y": 37
|
|
},
|
|
"id": 17,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "rate(process_cpu_seconds_total{job=\"apiserver\",instance=~\"$instance\", cluster=\"$cluster\"}[$__rate_interval])",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"title": "CPU usage",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"min": 0,
|
|
"show": true
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"aliasColors": {},
|
|
"bars": false,
|
|
"dashLength": 10,
|
|
"dashes": false,
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"fill": 1,
|
|
"fillGradient": 0,
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 16,
|
|
"y": 37
|
|
},
|
|
"id": 18,
|
|
"interval": "1m",
|
|
"legend": {
|
|
"alignAsTable": true,
|
|
"avg": false,
|
|
"current": false,
|
|
"max": false,
|
|
"min": false,
|
|
"rightSide": true,
|
|
"show": true,
|
|
"total": false,
|
|
"values": false
|
|
},
|
|
"lines": true,
|
|
"linewidth": 1,
|
|
"links": [],
|
|
"nullPointMode": "null",
|
|
"percentage": false,
|
|
"pointradius": 5,
|
|
"points": false,
|
|
"renderer": "flot",
|
|
"seriesOverrides": [],
|
|
"spaceLength": 10,
|
|
"stack": false,
|
|
"steppedLine": false,
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"uid": "$datasource"
|
|
},
|
|
"expr": "go_goroutines{job=\"apiserver\",instance=~\"$instance\", cluster=\"$cluster\"}",
|
|
"format": "time_series",
|
|
"intervalFactor": 2,
|
|
"legendFormat": "{{instance}}",
|
|
"refId": "A"
|
|
}
|
|
],
|
|
"thresholds": [],
|
|
"title": "Goroutines",
|
|
"tooltip": {
|
|
"shared": false,
|
|
"sort": 0,
|
|
"value_type": "individual"
|
|
},
|
|
"type": "graph",
|
|
"xaxis": {
|
|
"mode": "time",
|
|
"show": true,
|
|
"values": []
|
|
},
|
|
"yaxes": [
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"show": true
|
|
},
|
|
{
|
|
"format": "short",
|
|
"logBase": 1,
|
|
"show": true
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"refresh": "10s",
|
|
"schemaVersion": 36,
|
|
"style": "dark",
|
|
"tags": [
|
|
"kubernetes-mixin"
|
|
],
|
|
"templating": {
|
|
"list": [
|
|
{
|
|
"current": {
|
|
"selected": false,
|
|
"text": "default",
|
|
"value": "default"
|
|
},
|
|
"hide": 0,
|
|
"includeAll": false,
|
|
"label": "Data Source",
|
|
"multi": false,
|
|
"name": "datasource",
|
|
"options": [],
|
|
"query": "prometheus",
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"type": "datasource"
|
|
},
|
|
{
|
|
"current": {
|
|
"isNone": true,
|
|
"selected": false,
|
|
"text": "None",
|
|
"value": ""
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "$datasource"
|
|
},
|
|
"definition": "",
|
|
"hide": 2,
|
|
"includeAll": false,
|
|
"label": "cluster",
|
|
"multi": false,
|
|
"name": "cluster",
|
|
"options": [],
|
|
"query": {
|
|
"query": "label_values(up{job=\"apiserver\"}, cluster)",
|
|
"refId": "Prometheus-cluster-Variable-Query"
|
|
},
|
|
"refresh": 2,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 1,
|
|
"tagValuesQuery": "",
|
|
"tagsQuery": "",
|
|
"type": "query",
|
|
"useTags": false
|
|
},
|
|
{
|
|
"current": {
|
|
"selected": false,
|
|
"text": "All",
|
|
"value": "$__all"
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "$datasource"
|
|
},
|
|
"definition": "",
|
|
"hide": 0,
|
|
"includeAll": true,
|
|
"multi": false,
|
|
"name": "instance",
|
|
"options": [],
|
|
"query": {
|
|
"query": "label_values(up{job=\"apiserver\", cluster=\"$cluster\"}, instance)",
|
|
"refId": "Prometheus-instance-Variable-Query"
|
|
},
|
|
"refresh": 2,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 1,
|
|
"tagValuesQuery": "",
|
|
"tagsQuery": "",
|
|
"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": "utc",
|
|
"title": "Kubernetes / API server",
|
|
"uid": "09ec8aa1e996d6ffcd6817bbaff4db1b",
|
|
"version": 1,
|
|
"weekStart": ""
|
|
}
|