1105 lines
30 KiB
YAML
1105 lines
30 KiB
YAML
---
|
|
apiVersion: v1
|
|
data:
|
|
amd-gpu-dashboard.json: |-
|
|
{
|
|
"annotations": {
|
|
"list": [
|
|
{
|
|
"builtIn": 1,
|
|
"datasource": {
|
|
"type": "grafana",
|
|
"uid": "-- Grafana --"
|
|
},
|
|
"enable": true,
|
|
"hide": true,
|
|
"iconColor": "rgba(0, 211, 255, 1)",
|
|
"name": "Annotations & Alerts",
|
|
"type": "dashboard"
|
|
}
|
|
]
|
|
},
|
|
"editable": true,
|
|
"fiscalYearStartMonth": 0,
|
|
"graphTooltip": 0,
|
|
"id": 1,
|
|
"links": [],
|
|
"panels": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"gridPos": {
|
|
"h": 6,
|
|
"w": 4,
|
|
"x": 0,
|
|
"y": 0
|
|
},
|
|
"id": 10,
|
|
"options": {
|
|
"code": {
|
|
"language": "plaintext",
|
|
"showLineNumbers": false,
|
|
"showMiniMap": false
|
|
},
|
|
"content": "**GPU MODEL**\n## ${gpu_model}\n\n**GPU S/N**\n## ${gpu_serial_number}",
|
|
"mode": "markdown"
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"type": "text"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "watt"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 3,
|
|
"w": 5,
|
|
"x": 4,
|
|
"y": 0
|
|
},
|
|
"id": 6,
|
|
"options": {
|
|
"colorMode": "value",
|
|
"graphMode": "area",
|
|
"justifyMode": "auto",
|
|
"orientation": "auto",
|
|
"percentChangeColorMode": "standard",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"lastNotNull"
|
|
],
|
|
"fields": "/^gpu_power_usage$/",
|
|
"values": false
|
|
},
|
|
"showPercentChange": false,
|
|
"textMode": "auto",
|
|
"wideLayout": true
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_power_usage{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": true,
|
|
"legendFormat": "gpu_power_usage",
|
|
"range": false,
|
|
"refId": "gpu_power_usage",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "Power Usage (W)",
|
|
"transformations": [
|
|
{
|
|
"id": "merge",
|
|
"options": {}
|
|
},
|
|
{
|
|
"id": "reduce",
|
|
"options": {
|
|
"includeTimeField": false,
|
|
"mode": "reduceFields",
|
|
"reducers": [
|
|
"mean"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 6,
|
|
"w": 5,
|
|
"x": 9,
|
|
"y": 0
|
|
},
|
|
"id": 4,
|
|
"options": {
|
|
"minVizHeight": 75,
|
|
"minVizWidth": 75,
|
|
"orientation": "auto",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"lastNotNull"
|
|
],
|
|
"fields": "",
|
|
"values": false
|
|
},
|
|
"showThresholdLabels": false,
|
|
"showThresholdMarkers": true,
|
|
"sizing": "auto"
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_used_vram{gpu_uuid=~\"$gpu\", job=\"$scope\"} / on(gpu_uuid) gpu_total_vram{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": true,
|
|
"legendFormat": "gpu_vram_utilization",
|
|
"range": false,
|
|
"refId": "gpu_used_vram",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "Memory Utilization %",
|
|
"transformations": [
|
|
{
|
|
"id": "joinByField",
|
|
"options": {
|
|
"byField": "Time",
|
|
"mode": "outer"
|
|
}
|
|
},
|
|
{
|
|
"id": "reduce",
|
|
"options": {
|
|
"reducers": [
|
|
"mean"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"type": "gauge"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 6,
|
|
"w": 5,
|
|
"x": 14,
|
|
"y": 0
|
|
},
|
|
"id": 12,
|
|
"options": {
|
|
"minVizHeight": 75,
|
|
"minVizWidth": 75,
|
|
"orientation": "auto",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"lastNotNull"
|
|
],
|
|
"fields": "/^gpu_gfx_activity$/",
|
|
"values": false
|
|
},
|
|
"showThresholdLabels": false,
|
|
"showThresholdMarkers": true,
|
|
"sizing": "auto"
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_gfx_activity{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": true,
|
|
"legendFormat": "gpu_gfx_activity",
|
|
"range": false,
|
|
"refId": "gpu_gfx_activity",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "GPU Utilization %",
|
|
"transformations": [
|
|
{
|
|
"id": "merge",
|
|
"options": {}
|
|
},
|
|
{
|
|
"id": "reduce",
|
|
"options": {
|
|
"includeTimeField": false,
|
|
"mode": "reduceFields",
|
|
"reducers": [
|
|
"mean"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"type": "gauge"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "celsius"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 6,
|
|
"w": 5,
|
|
"x": 19,
|
|
"y": 0
|
|
},
|
|
"id": 9,
|
|
"options": {
|
|
"colorMode": "value",
|
|
"graphMode": "area",
|
|
"justifyMode": "auto",
|
|
"orientation": "auto",
|
|
"percentChangeColorMode": "standard",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"lastNotNull"
|
|
],
|
|
"fields": "",
|
|
"values": false
|
|
},
|
|
"showPercentChange": false,
|
|
"textMode": "auto",
|
|
"wideLayout": true
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_edge_temperature{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": true,
|
|
"legendFormat": "gpu_edge_temperature",
|
|
"range": false,
|
|
"refId": "gpu_edge_temperature",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "GPU Temperature °C",
|
|
"transformations": [
|
|
{
|
|
"id": "merge",
|
|
"options": {}
|
|
},
|
|
{
|
|
"id": "reduce",
|
|
"options": {
|
|
"includeTimeField": false,
|
|
"labelsToFields": false,
|
|
"mode": "seriesToRows",
|
|
"reducers": [
|
|
"mean"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 3,
|
|
"w": 5,
|
|
"x": 4,
|
|
"y": 3
|
|
},
|
|
"id": 13,
|
|
"options": {
|
|
"colorMode": "value",
|
|
"graphMode": "area",
|
|
"justifyMode": "auto",
|
|
"orientation": "auto",
|
|
"percentChangeColorMode": "standard",
|
|
"reduceOptions": {
|
|
"calcs": [
|
|
"lastNotNull"
|
|
],
|
|
"fields": "/^gpu_used_vram$/",
|
|
"values": false
|
|
},
|
|
"showPercentChange": false,
|
|
"textMode": "auto",
|
|
"wideLayout": true
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_used_vram{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": true,
|
|
"legendFormat": "gpu_used_vram",
|
|
"range": false,
|
|
"refId": "gpu_used_vram",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "VRAM Usage (MB)",
|
|
"transformations": [
|
|
{
|
|
"id": "merge",
|
|
"options": {}
|
|
},
|
|
{
|
|
"id": "reduce",
|
|
"options": {
|
|
"includeTimeField": false,
|
|
"mode": "reduceFields",
|
|
"reducers": [
|
|
"mean"
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"type": "stat"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"axisBorderShow": false,
|
|
"axisCenteredZero": false,
|
|
"axisColorMode": "text",
|
|
"axisLabel": "",
|
|
"axisPlacement": "auto",
|
|
"barAlignment": 0,
|
|
"drawStyle": "line",
|
|
"fillOpacity": 0,
|
|
"gradientMode": "none",
|
|
"hideFrom": {
|
|
"legend": false,
|
|
"tooltip": false,
|
|
"viz": false
|
|
},
|
|
"insertNulls": false,
|
|
"lineInterpolation": "linear",
|
|
"lineWidth": 1,
|
|
"pointSize": 5,
|
|
"scaleDistribution": {
|
|
"type": "linear"
|
|
},
|
|
"showPoints": "auto",
|
|
"spanNulls": false,
|
|
"stacking": {
|
|
"group": "A",
|
|
"mode": "none"
|
|
},
|
|
"thresholdsStyle": {
|
|
"mode": "off"
|
|
}
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 0,
|
|
"y": 6
|
|
},
|
|
"id": 3,
|
|
"options": {
|
|
"legend": {
|
|
"calcs": [],
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true
|
|
},
|
|
"tooltip": {
|
|
"mode": "single",
|
|
"sort": "none"
|
|
}
|
|
},
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_gfx_activity{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"includeNullMetadata": true,
|
|
"instant": false,
|
|
"legendFormat": "{{gpu_uuid}}",
|
|
"range": true,
|
|
"refId": "gpu_gfx_activity",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "GPU Utilization %",
|
|
"type": "timeseries"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"axisBorderShow": false,
|
|
"axisCenteredZero": false,
|
|
"axisColorMode": "text",
|
|
"axisLabel": "",
|
|
"axisPlacement": "auto",
|
|
"axisSoftMax": 1,
|
|
"axisSoftMin": 0,
|
|
"barAlignment": 0,
|
|
"drawStyle": "line",
|
|
"fillOpacity": 0,
|
|
"gradientMode": "none",
|
|
"hideFrom": {
|
|
"legend": false,
|
|
"tooltip": false,
|
|
"viz": false
|
|
},
|
|
"insertNulls": false,
|
|
"lineInterpolation": "linear",
|
|
"lineWidth": 1,
|
|
"pointSize": 5,
|
|
"scaleDistribution": {
|
|
"type": "linear"
|
|
},
|
|
"showPoints": "auto",
|
|
"spanNulls": false,
|
|
"stacking": {
|
|
"group": "A",
|
|
"mode": "none"
|
|
},
|
|
"thresholdsStyle": {
|
|
"mode": "off"
|
|
}
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "percentunit"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 8,
|
|
"y": 6
|
|
},
|
|
"id": 2,
|
|
"options": {
|
|
"legend": {
|
|
"calcs": [],
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true
|
|
},
|
|
"tooltip": {
|
|
"mode": "single",
|
|
"sort": "none"
|
|
}
|
|
},
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_used_vram{gpu_uuid=~\"$gpu\", job=\"$scope\"} / on(gpu_uuid) gpu_total_vram{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"includeNullMetadata": true,
|
|
"instant": false,
|
|
"legendFormat": "__auto",
|
|
"range": true,
|
|
"refId": "gpu_used_vram",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "Memory Utilization %",
|
|
"type": "timeseries"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "palette-classic"
|
|
},
|
|
"custom": {
|
|
"axisBorderShow": false,
|
|
"axisCenteredZero": false,
|
|
"axisColorMode": "text",
|
|
"axisLabel": "",
|
|
"axisPlacement": "auto",
|
|
"barAlignment": 0,
|
|
"drawStyle": "line",
|
|
"fillOpacity": 0,
|
|
"gradientMode": "none",
|
|
"hideFrom": {
|
|
"legend": false,
|
|
"tooltip": false,
|
|
"viz": false
|
|
},
|
|
"insertNulls": false,
|
|
"lineInterpolation": "linear",
|
|
"lineWidth": 1,
|
|
"pointSize": 5,
|
|
"scaleDistribution": {
|
|
"type": "linear"
|
|
},
|
|
"showPoints": "auto",
|
|
"spanNulls": false,
|
|
"stacking": {
|
|
"group": "A",
|
|
"mode": "none"
|
|
},
|
|
"thresholdsStyle": {
|
|
"mode": "off"
|
|
}
|
|
},
|
|
"mappings": [],
|
|
"min": 0,
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
},
|
|
"unit": "celsius"
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 8,
|
|
"x": 16,
|
|
"y": 6
|
|
},
|
|
"id": 11,
|
|
"options": {
|
|
"legend": {
|
|
"calcs": [],
|
|
"displayMode": "list",
|
|
"placement": "bottom",
|
|
"showLegend": true
|
|
},
|
|
"tooltip": {
|
|
"mode": "single",
|
|
"sort": "none"
|
|
}
|
|
},
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"exemplar": false,
|
|
"expr": "gpu_edge_temperature{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "time_series",
|
|
"fullMetaSearch": false,
|
|
"includeNullMetadata": true,
|
|
"instant": false,
|
|
"legendFormat": "{{gpu_uuid}}",
|
|
"range": true,
|
|
"refId": "gpu_edge_temperature",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "GPU Temperature °C",
|
|
"type": "timeseries"
|
|
},
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"fieldConfig": {
|
|
"defaults": {
|
|
"color": {
|
|
"mode": "thresholds"
|
|
},
|
|
"custom": {
|
|
"align": "auto",
|
|
"cellOptions": {
|
|
"type": "auto",
|
|
"wrapText": false
|
|
},
|
|
"filterable": true,
|
|
"inspect": false
|
|
},
|
|
"mappings": [],
|
|
"thresholds": {
|
|
"mode": "absolute",
|
|
"steps": [
|
|
{
|
|
"color": "green",
|
|
"value": null
|
|
},
|
|
{
|
|
"color": "red",
|
|
"value": 80
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"overrides": []
|
|
},
|
|
"gridPos": {
|
|
"h": 7,
|
|
"w": 24,
|
|
"x": 0,
|
|
"y": 13
|
|
},
|
|
"id": 14,
|
|
"options": {
|
|
"cellHeight": "sm",
|
|
"footer": {
|
|
"countRows": false,
|
|
"fields": "",
|
|
"reducer": [
|
|
"sum"
|
|
],
|
|
"show": false
|
|
},
|
|
"showHeader": true
|
|
},
|
|
"pluginVersion": "11.1.4",
|
|
"targets": [
|
|
{
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"disableTextWrap": false,
|
|
"editorMode": "builder",
|
|
"expr": "{gpu_uuid=~\"$gpu\", job=\"$scope\"}",
|
|
"format": "table",
|
|
"fullMetaSearch": false,
|
|
"hide": false,
|
|
"includeNullMetadata": true,
|
|
"instant": false,
|
|
"legendFormat": "__auto",
|
|
"range": true,
|
|
"refId": "A",
|
|
"useBackend": false
|
|
}
|
|
],
|
|
"title": "Raw data",
|
|
"type": "table"
|
|
}
|
|
],
|
|
"schemaVersion": 39,
|
|
"tags": [],
|
|
"templating": {
|
|
"list": [
|
|
{
|
|
"current": {
|
|
"selected": false,
|
|
"text": "kubernetes-pods",
|
|
"value": "kubernetes-pods"
|
|
},
|
|
"description": "Filter variable for job scope",
|
|
"hide": 0,
|
|
"includeAll": false,
|
|
"label": "Scope",
|
|
"multi": false,
|
|
"name": "scope",
|
|
"options": [
|
|
{
|
|
"selected": true,
|
|
"text": "kubernetes-pods",
|
|
"value": "kubernetes-pods"
|
|
},
|
|
{
|
|
"selected": false,
|
|
"text": "kubernetes-service-endpoints",
|
|
"value": "kubernetes-service-endpoints"
|
|
}
|
|
],
|
|
"query": "kubernetes-pods,kubernetes-service-endpoints",
|
|
"queryValue": "",
|
|
"skipUrlSync": false,
|
|
"type": "custom"
|
|
},
|
|
{
|
|
"current": {
|
|
"selected": false,
|
|
"text": "atlpkuc4app06",
|
|
"value": "atlpkuc4app06"
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"definition": "label_values(node)",
|
|
"hide": 0,
|
|
"includeAll": false,
|
|
"label": "Node",
|
|
"multi": false,
|
|
"name": "node",
|
|
"options": [],
|
|
"query": {
|
|
"qryType": 1,
|
|
"query": "label_values(node)",
|
|
"refId": "PrometheusVariableQueryEditor-VariableQuery"
|
|
},
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 0,
|
|
"type": "query"
|
|
},
|
|
{
|
|
"current": {
|
|
"selected": true,
|
|
"text": "All",
|
|
"value": "$__all"
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"definition": "label_values({node=\"$node\"},gpu_uuid)",
|
|
"hide": 0,
|
|
"includeAll": true,
|
|
"label": "GPU",
|
|
"multi": false,
|
|
"name": "gpu",
|
|
"options": [],
|
|
"query": {
|
|
"qryType": 1,
|
|
"query": "label_values({node=\"$node\"},gpu_uuid)",
|
|
"refId": "PrometheusVariableQueryEditor-VariableQuery"
|
|
},
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 0,
|
|
"type": "query"
|
|
},
|
|
{
|
|
"current": {
|
|
"isNone": true,
|
|
"selected": false,
|
|
"text": "None",
|
|
"value": ""
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"definition": "label_values({gpu_uuid=\"$gpu\"},amd_com_gpu_family)",
|
|
"hide": 2,
|
|
"includeAll": false,
|
|
"label": "GPU Family",
|
|
"multi": false,
|
|
"name": "gpu_family",
|
|
"options": [],
|
|
"query": {
|
|
"qryType": 1,
|
|
"query": "label_values({gpu_uuid=\"$gpu\"},amd_com_gpu_family)",
|
|
"refId": "PrometheusVariableQueryEditor-VariableQuery"
|
|
},
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 0,
|
|
"type": "query"
|
|
},
|
|
{
|
|
"current": {
|
|
"selected": true,
|
|
"text": [
|
|
"102-D65210-00"
|
|
],
|
|
"value": [
|
|
"102-D65210-00"
|
|
]
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"definition": "label_values({gpu_uuid=~\"$gpu\"},card_model)",
|
|
"hide": 2,
|
|
"includeAll": false,
|
|
"label": "GPU Model",
|
|
"multi": true,
|
|
"name": "gpu_model",
|
|
"options": [],
|
|
"query": {
|
|
"qryType": 1,
|
|
"query": "label_values({gpu_uuid=~\"$gpu\"},card_model)",
|
|
"refId": "PrometheusVariableQueryEditor-VariableQuery"
|
|
},
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 0,
|
|
"type": "query"
|
|
},
|
|
{
|
|
"current": {
|
|
"selected": true,
|
|
"text": [
|
|
"692324000620"
|
|
],
|
|
"value": [
|
|
"692324000620"
|
|
]
|
|
},
|
|
"datasource": {
|
|
"type": "prometheus",
|
|
"uid": "prometheus"
|
|
},
|
|
"definition": "label_values({gpu_uuid=~\"$gpu\"},serial_number)",
|
|
"hide": 2,
|
|
"includeAll": false,
|
|
"label": "GPU Serial Number",
|
|
"multi": true,
|
|
"name": "gpu_serial_number",
|
|
"options": [],
|
|
"query": {
|
|
"qryType": 1,
|
|
"query": "label_values({gpu_uuid=~\"$gpu\"},serial_number)",
|
|
"refId": "PrometheusVariableQueryEditor-VariableQuery"
|
|
},
|
|
"refresh": 1,
|
|
"regex": "",
|
|
"skipUrlSync": false,
|
|
"sort": 0,
|
|
"type": "query"
|
|
}
|
|
]
|
|
},
|
|
"time": {
|
|
"from": "now-1h",
|
|
"to": "now"
|
|
},
|
|
"timepicker": {},
|
|
"timezone": "browser",
|
|
"title": "AMD GPU Metrics",
|
|
"uid": "ee6sumbzuzl6oc",
|
|
"version": 25,
|
|
"weekStart": ""
|
|
}
|
|
kind: ConfigMap
|
|
metadata:
|
|
annotations:
|
|
grafana_folder: common
|
|
labels:
|
|
grafana_dashboard: "1"
|
|
name: lgtm-amd-gpu-dashboard
|
|
namespace: otel-lgtm-stack
|