652b02b53f
Three dashboards over metrics already in VictoriaMetrics: the swarm controller's job graph and forge push/commit counters, the forge process's runtime, and the OTEL collectors' self-telemetry.
491 lines
12 KiB
JSON
491 lines
12 KiB
JSON
{
|
|
"title": "hyperhive \u00b7 swarm controller (jobq + vcs)",
|
|
"uid": "hyperhive-swarm-controller",
|
|
"description": "The swarm controller's job graph and the forge activity it sees.",
|
|
"editable": true,
|
|
"refresh": "1m",
|
|
"schemaVersion": 39,
|
|
"tags": ["hyperhive", "swarm", "controller", "jobq", "vcs"],
|
|
"time": {
|
|
"from": "now-24h",
|
|
"to": "now"
|
|
},
|
|
"timezone": "utc",
|
|
"version": 1,
|
|
"panels": [
|
|
{
|
|
"id": 1,
|
|
"type": "stat",
|
|
"title": "Running nodes",
|
|
"description": "Job nodes executing now, at any depth.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 4,
|
|
"w": 6,
|
|
"x": 0,
|
|
"y": 0
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum({__name__=\"jobq.nodes\", \"service.name\"=\"swarm-controller\", state=\"Running\"}) or vector(0)",
|
|
"instant": true,
|
|
"legendFormat": "running"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"decimals": 0,
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "text",
|
|
"value": null
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"graphMode": "none",
|
|
"colorMode": "value",
|
|
"textMode": "value",
|
|
"justifyMode": "auto",
|
|
"reduceOptions": {
|
|
"calcs": ["lastNotNull"],
|
|
"fields": "",
|
|
"values": false
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 2,
|
|
"type": "stat",
|
|
"title": "Pending nodes",
|
|
"description": "Job nodes waiting on dependencies or resources.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 4,
|
|
"w": 6,
|
|
"x": 6,
|
|
"y": 0
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum({__name__=\"jobq.nodes\", \"service.name\"=\"swarm-controller\", state=\"Pending\"}) or vector(0)",
|
|
"instant": true,
|
|
"legendFormat": "pending"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"decimals": 0,
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "text",
|
|
"value": null
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"graphMode": "none",
|
|
"colorMode": "value",
|
|
"textMode": "value",
|
|
"justifyMode": "auto",
|
|
"reduceOptions": {
|
|
"calcs": ["lastNotNull"],
|
|
"fields": "",
|
|
"values": false
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 3,
|
|
"type": "stat",
|
|
"title": "Failed root groups",
|
|
"description": "Root job groups in the Failed state.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 4,
|
|
"w": 6,
|
|
"x": 12,
|
|
"y": 0
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum({__name__=\"jobq.roots\", \"service.name\"=\"swarm-controller\", state=\"Failed\"}) or vector(0)",
|
|
"instant": true,
|
|
"legendFormat": "failed"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"decimals": 0,
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 1
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"graphMode": "none",
|
|
"colorMode": "value",
|
|
"textMode": "value",
|
|
"justifyMode": "auto",
|
|
"reduceOptions": {
|
|
"calcs": ["lastNotNull"],
|
|
"fields": "",
|
|
"values": false
|
|
}
|
|
},
|
|
"x-zero-is-healthy": true
|
|
},
|
|
{
|
|
"id": 4,
|
|
"type": "stat",
|
|
"title": "Pushes (range)",
|
|
"description": "Push deliveries seen in the selected range.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 4,
|
|
"w": 6,
|
|
"x": 18,
|
|
"y": 0
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum(increase({__name__=\"hyperhive.vcs.push.count\", \"service.name\"=\"swarm-controller\"}[$__range])) or vector(0)",
|
|
"instant": true,
|
|
"legendFormat": "pushes"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"decimals": 0,
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "text",
|
|
"value": null
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"graphMode": "none",
|
|
"colorMode": "value",
|
|
"textMode": "value",
|
|
"justifyMode": "auto",
|
|
"reduceOptions": {
|
|
"calcs": ["lastNotNull"],
|
|
"fields": "",
|
|
"values": false
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 5,
|
|
"type": "timeseries",
|
|
"title": "Nodes by state",
|
|
"description": "Job nodes at any depth, per lifecycle state.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 8,
|
|
"w": 12,
|
|
"x": 0,
|
|
"y": 4
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum by (state) ({__name__=\"jobq.nodes\", \"service.name\"=\"swarm-controller\"})",
|
|
"legendFormat": "{{state}}"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"drawStyle": "line",
|
|
"lineWidth": 1,
|
|
"fillOpacity": 10,
|
|
"showPoints": "never",
|
|
"spanNulls": false,
|
|
"axisSoftMin": 0
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"legend": {
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true,
|
|
"calcs": []
|
|
},
|
|
"tooltip": {
|
|
"mode": "multi",
|
|
"sort": "desc"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 6,
|
|
"type": "timeseries",
|
|
"title": "Root groups by state",
|
|
"description": "Root job groups, per lifecycle state.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 8,
|
|
"w": 12,
|
|
"x": 12,
|
|
"y": 4
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum by (state) ({__name__=\"jobq.roots\", \"service.name\"=\"swarm-controller\"})",
|
|
"legendFormat": "{{state}}"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"drawStyle": "line",
|
|
"lineWidth": 1,
|
|
"fillOpacity": 10,
|
|
"showPoints": "never",
|
|
"spanNulls": false,
|
|
"axisSoftMin": 0
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"legend": {
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true,
|
|
"calcs": []
|
|
},
|
|
"tooltip": {
|
|
"mode": "multi",
|
|
"sort": "desc"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 7,
|
|
"type": "timeseries",
|
|
"title": "Pushes and commits",
|
|
"description": "Push deliveries and commits per minute.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 8,
|
|
"w": 12,
|
|
"x": 0,
|
|
"y": 12
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum(rate({__name__=\"hyperhive.vcs.push.count\", \"service.name\"=\"swarm-controller\"}[$__rate_interval])) * 60",
|
|
"legendFormat": "pushes"
|
|
},
|
|
{
|
|
"refId": "B",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sum(rate({__name__=\"hyperhive.vcs.commit.count\", \"service.name\"=\"swarm-controller\"}[$__rate_interval])) * 60",
|
|
"legendFormat": "commits"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"drawStyle": "line",
|
|
"lineWidth": 1,
|
|
"fillOpacity": 10,
|
|
"showPoints": "never",
|
|
"spanNulls": false,
|
|
"axisSoftMin": 0
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"legend": {
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true,
|
|
"calcs": []
|
|
},
|
|
"tooltip": {
|
|
"mode": "multi",
|
|
"sort": "desc"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"id": 8,
|
|
"type": "timeseries",
|
|
"title": "Commits by repo",
|
|
"description": "Commits per hour, per repository.",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"gridPos": {
|
|
"h": 8,
|
|
"w": 12,
|
|
"x": 12,
|
|
"y": 12
|
|
},
|
|
"targets": [
|
|
{
|
|
"refId": "A",
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "@datasourceUid@"
|
|
},
|
|
"expr": "sort_desc(sum by (repo) (rate({__name__=\"hyperhive.vcs.commit.count\", \"service.name\"=\"swarm-controller\"}[$__rate_interval])) * 3600)",
|
|
"legendFormat": "{{repo}}"
|
|
}
|
|
],
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"unit": "short",
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"drawStyle": "line",
|
|
"lineWidth": 1,
|
|
"fillOpacity": 10,
|
|
"showPoints": "never",
|
|
"spanNulls": false,
|
|
"axisSoftMin": 0
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"options": {
|
|
"legend": {
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true,
|
|
"calcs": []
|
|
},
|
|
"tooltip": {
|
|
"mode": "multi",
|
|
"sort": "desc"
|
|
}
|
|
}
|
|
}
|
|
]
|
|
}
|