incus-contrib/incusos/observability-dashboards/host-resources.json

430 lines
13 KiB
JSON

{
"uid": "incus-host-resources",
"title": "Host Resources (IncusOS Nodes)",
"description": "Physical host monitoring for IncusOS cluster nodes: CPU, memory, disk, and network from node_exporter metrics.",
"tags": ["incusos", "host", "node-exporter"],
"timezone": "browser",
"schemaVersion": 39,
"version": 1,
"refresh": "30s",
"time": {
"from": "now-1h",
"to": "now"
},
"fiscalYearStartMonth": 0,
"liveNow": false,
"editable": true,
"graphTooltip": 1,
"templating": {
"list": []
},
"annotations": {
"list": [
{
"builtIn": 1,
"datasource": {
"type": "grafana",
"uid": "-- Grafana --"
},
"enable": true,
"hide": true,
"iconColor": "rgba(0, 211, 255, 1)",
"name": "Annotations & Alerts",
"type": "dashboard"
}
]
},
"panels": [
{
"type": "row",
"title": "CPU",
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
"id": 1,
"panels": []
},
{
"type": "timeseries",
"title": "CPU Usage % by Node",
"description": "Overall CPU utilization per host node, calculated as 1 minus the average idle CPU rate.",
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 1 },
"id": 2,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "(1 - avg by (instance) (rate(node_cpu_seconds_total{mode=\"idle\"}[$__rate_interval]))) * 100",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 2,
"fillOpacity": 25,
"gradientMode": "scheme",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "line" }
},
"unit": "percent",
"min": 0,
"max": 100,
"color": { "mode": "continuous-GrYlRd" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null },
{ "color": "yellow", "value": 70 },
{ "color": "red", "value": 90 }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "table", "placement": "right", "calcs": ["mean", "max", "lastNotNull"] }
}
},
{
"type": "row",
"title": "Memory",
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 9 },
"id": 3,
"panels": []
},
{
"type": "timeseries",
"title": "Memory Usage % by Node",
"description": "Memory utilization percentage per host node, calculated from available vs total memory.",
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 10 },
"id": 4,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "(1 - node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 2,
"fillOpacity": 25,
"gradientMode": "scheme",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "line" }
},
"unit": "percent",
"min": 0,
"max": 100,
"color": { "mode": "continuous-GrYlRd" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null },
{ "color": "yellow", "value": 70 },
{ "color": "red", "value": 90 }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "table", "placement": "right", "calcs": ["mean", "max", "lastNotNull"] }
}
},
{
"type": "row",
"title": "Disk",
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 18 },
"id": 5,
"panels": []
},
{
"type": "timeseries",
"title": "Root Filesystem Available Space",
"description": "Available disk space on the root filesystem (mountpoint='/') per host node.",
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 19 },
"id": 6,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "node_filesystem_avail_bytes{mountpoint=\"/\"}",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 1,
"fillOpacity": 15,
"gradientMode": "none",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "off" }
},
"unit": "bytes",
"color": { "mode": "palette-classic" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "table", "placement": "right", "calcs": ["lastNotNull", "min"] }
}
},
{
"type": "timeseries",
"title": "Disk Read Rate by Node",
"description": "Disk read throughput (bytes/sec) aggregated per host node.",
"gridPos": { "h": 8, "w": 6, "x": 12, "y": 19 },
"id": 7,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "sum by (instance) (rate(node_disk_read_bytes_total[$__rate_interval]))",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 1,
"fillOpacity": 15,
"gradientMode": "none",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "off" }
},
"unit": "Bps",
"color": { "mode": "palette-classic" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "list", "placement": "bottom", "calcs": [] }
}
},
{
"type": "timeseries",
"title": "Disk Write Rate by Node",
"description": "Disk write throughput (bytes/sec) aggregated per host node.",
"gridPos": { "h": 8, "w": 6, "x": 18, "y": 19 },
"id": 8,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "sum by (instance) (rate(node_disk_written_bytes_total[$__rate_interval]))",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 1,
"fillOpacity": 15,
"gradientMode": "none",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "off" }
},
"unit": "Bps",
"color": { "mode": "palette-classic" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "list", "placement": "bottom", "calcs": [] }
}
},
{
"type": "row",
"title": "Network",
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 27 },
"id": 9,
"panels": []
},
{
"type": "timeseries",
"title": "Network Receive Rate by Device & Node",
"description": "Inbound network throughput per network device and host node.",
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 28 },
"id": 10,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "rate(node_network_receive_bytes_total{device!~\"lo|veth.*|br-.*|docker.*\"}[$__rate_interval])",
"legendFormat": "{{instance}} - {{device}} (rx)",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 1,
"fillOpacity": 15,
"gradientMode": "none",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "off" }
},
"unit": "Bps",
"color": { "mode": "palette-classic" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "table", "placement": "right", "calcs": ["mean", "max"] }
}
},
{
"type": "timeseries",
"title": "Network Transmit Rate by Device & Node",
"description": "Outbound network throughput per network device and host node.",
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 28 },
"id": 11,
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"targets": [
{
"expr": "rate(node_network_transmit_bytes_total{device!~\"lo|veth.*|br-.*|docker.*\"}[$__rate_interval])",
"legendFormat": "{{instance}} - {{device}} (tx)",
"refId": "A"
}
],
"fieldConfig": {
"defaults": {
"custom": {
"drawStyle": "line",
"lineInterpolation": "smooth",
"lineWidth": 1,
"fillOpacity": 15,
"gradientMode": "none",
"spanNulls": false,
"showPoints": "never",
"pointSize": 5,
"stacking": { "mode": "none", "group": "A" },
"axisPlacement": "auto",
"axisLabel": "",
"scaleDistribution": { "type": "linear" },
"thresholdsStyle": { "mode": "off" }
},
"unit": "Bps",
"color": { "mode": "palette-classic" },
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null }
]
}
},
"overrides": []
},
"options": {
"tooltip": { "mode": "multi", "sort": "desc" },
"legend": { "displayMode": "table", "placement": "right", "calcs": ["mean", "max"] }
}
}
]
}