maas-base/docker-compose/grafana/grafana_dashboards/gpustack-worker.json

{
  "annotations": {
    "list": [
      {
        "builtIn": 1,
        "datasource": {
          "type": "datasource",
          "uid": "grafana"
        },
        "enable": true,
        "hide": true,
        "iconColor": "rgba(0, 211, 255, 1)",
        "name": "Annotations & Alerts",
        "type": "dashboard"
      }
    ]
  },
  "editable": true,
  "fiscalYearStartMonth": 0,
  "graphTooltip": 0,
  "id": 11,
  "links": [],
  "panels": [
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "decimals": 0,
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "blue",
                "value": 0
              }
            ]
          }
        },
        "overrides": [
          {
            "matcher": {
              "id": "byName",
              "options": "Total Workers"
            },
            "properties": [
              {
                "id": "color",
                "value": {
                  "fixedColor": "blue",
                  "mode": "fixed"
                }
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "Total GPUs"
            },
            "properties": [
              {
                "id": "color",
                "value": {
                  "fixedColor": "yellow",
                  "mode": "fixed"
                }
              }
            ]
          }
        ]
      },
      "gridPos": {
        "h": 4,
        "w": 9,
        "x": 0,
        "y": 0
      },
      "id": 74,
      "options": {
        "colorMode": "background",
        "graphMode": "none",
        "justifyMode": "auto",
        "orientation": "auto",
        "percentChangeColorMode": "standard",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showPercentChange": false,
        "textMode": "auto",
        "wideLayout": true
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "expr": "count(gpustack:worker_node_os_info{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "hide": false,
          "instant": false,
          "legendFormat": "Total Workers",
          "range": true,
          "refId": "E"
        },
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "expr": "count(gpustack:worker_node_gpu_cores{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "Total GPUs",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Summary",
      "type": "stat"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 5,
        "x": 9,
        "y": 0
      },
      "id": 89,
      "options": {
        "minVizHeight": 75,
        "minVizWidth": 75,
        "orientation": "horizontal",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showThresholdLabels": false,
        "showThresholdMarkers": true,
        "sizing": "auto"
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "avg(gpustack:worker_node_gpu_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Average GPU Utilization",
      "type": "gauge"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 5,
        "x": 14,
        "y": 0
      },
      "id": 90,
      "options": {
        "minVizHeight": 75,
        "minVizWidth": 75,
        "orientation": "horizontal",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showThresholdLabels": false,
        "showThresholdMarkers": true,
        "sizing": "auto"
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "avg(gpustack:worker_node_gram_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Average VRAM Utilization",
      "type": "gauge"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "celsius"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 5,
        "x": 19,
        "y": 0
      },
      "id": 82,
      "options": {
        "minVizHeight": 75,
        "minVizWidth": 75,
        "orientation": "horizontal",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showThresholdLabels": false,
        "showThresholdMarkers": true,
        "sizing": "auto"
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "avg(gpustack:worker_node_gpu_temperature_celsius{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Average GPU Temperature",
      "type": "gauge"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "bytes"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 4,
        "w": 3,
        "x": 0,
        "y": 4
      },
      "id": 95,
      "options": {
        "colorMode": "none",
        "graphMode": "none",
        "justifyMode": "auto",
        "orientation": "horizontal",
        "percentChangeColorMode": "standard",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showPercentChange": false,
        "textMode": "auto",
        "wideLayout": true
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "sum(gpustack:worker_node_gram_total_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "VRAM Size",
      "type": "stat"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "bytes"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 4,
        "w": 3,
        "x": 3,
        "y": 4
      },
      "id": 93,
      "options": {
        "colorMode": "value",
        "graphMode": "none",
        "justifyMode": "auto",
        "orientation": "horizontal",
        "percentChangeColorMode": "standard",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showPercentChange": false,
        "textMode": "auto",
        "wideLayout": true
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "avg(gpustack:worker_node_memory_total_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Memory Size",
      "type": "stat"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "max": 90,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "none"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 4,
        "w": 3,
        "x": 6,
        "y": 4
      },
      "id": 94,
      "options": {
        "colorMode": "value",
        "graphMode": "none",
        "justifyMode": "auto",
        "orientation": "horizontal",
        "percentChangeColorMode": "standard",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showPercentChange": false,
        "textMode": "auto",
        "wideLayout": true
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "P1809F7CD0C75ACF3"
          },
          "editorMode": "code",
          "expr": "avg(gpustack:worker_node_cpu_cores{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"})",
          "instant": false,
          "legendFormat": "__auto",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "CPU Cores",
      "type": "stat"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "description": "",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "custom": {
            "align": "center",
            "cellOptions": {
              "type": "auto"
            },
            "footer": {
              "reducers": []
            },
            "inspect": false
          },
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          }
        },
        "overrides": [
          {
            "matcher": {
              "id": "byName",
              "options": "Cluster"
            },
            "properties": [
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "color-background"
                }
              },
              {
                "id": "custom.width",
                "value": 173
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "GPU VRAM Util Rate"
            },
            "properties": [
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "color-background"
                }
              },
              {
                "id": "color",
                "value": {
                  "fixedColor": "blue",
                  "mode": "fixed"
                }
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "Worker"
            },
            "properties": [
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "color-text"
                }
              },
              {
                "id": "color",
                "value": {
                  "fixedColor": "blue",
                  "mode": "fixed"
                }
              },
              {
                "id": "custom.width",
                "value": 144
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "Chip Index"
            },
            "properties": [
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "color-text"
                }
              },
              {
                "id": "color",
                "value": {
                  "fixedColor": "blue",
                  "mode": "fixed"
                }
              },
              {
                "id": "custom.width",
                "value": 116
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "GPU Index"
            },
            "properties": [
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "color-text"
                }
              },
              {
                "id": "color",
                "value": {
                  "fixedColor": "blue",
                  "mode": "fixed"
                }
              },
              {
                "id": "custom.width",
                "value": 90
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "GPU VRAM Utilization"
            },
            "properties": [
              {
                "id": "unit",
                "value": "percent"
              },
              {
                "id": "decimals",
                "value": 1
              },
              {
                "id": "custom.cellOptions",
                "value": {
                  "type": "gauge"
                }
              },
              {
                "id": "color",
                "value": {
                  "mode": "continuous-GrYlRd"
                }
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "GPU Utilization"
            },
            "properties": [
              {
                "id": "unit",
                "value": "percent"
              },
              {
                "id": "decimals",
                "value": 1
              },
              {
                "id": "custom.cellOptions",
                "value": {
                  "mode": "gradient",
                  "type": "gauge",
                  "valueDisplayMode": "text"
                }
              },
              {
                "id": "color",
                "value": {
                  "mode": "continuous-GrYlRd"
                }
              },
              {
                "id": "custom.width",
                "value": 200
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "GPU Name"
            },
            "properties": [
              {
                "id": "custom.width",
                "value": 279
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "Chip Index"
            },
            "properties": [
              {
                "id": "custom.width",
                "value": 102
              }
            ]
          },
          {
            "matcher": {
              "id": "byName",
              "options": "Temperature"
            },
            "properties": [
              {
                "id": "custom.width",
                "value": 117
              }
            ]
          }
        ]
      },
      "gridPos": {
        "h": 7,
        "w": 24,
        "x": 0,
        "y": 8
      },
      "id": 66,
      "options": {
        "cellHeight": "sm",
        "frameIndex": 1,
        "showHeader": true,
        "sortBy": []
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "gpustack:worker_node_gpu_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"} ",
          "format": "table",
          "instant": true,
          "legendFormat": "__auto",
          "range": false,
          "refId": "A"
        },
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "gpustack:worker_node_gram_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"} ",
          "format": "table",
          "hide": false,
          "instant": true,
          "legendFormat": "__auto",
          "range": false,
          "refId": "B"
        },
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "gpustack:worker_node_gpu_temperature_celsius{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"} ",
          "format": "table",
          "hide": false,
          "instant": true,
          "legendFormat": "__auto",
          "range": false,
          "refId": "C"
        }
      ],
      "title": "GPUs",
      "transformations": [
        {
          "id": "concatenate",
          "options": {
            "frameNameLabel": "frame",
            "frameNameMode": "field"
          }
        },
        {
          "id": "organize",
          "options": {
            "excludeByName": {
              "Time": true,
              "__name__ 1": true,
              "__name__ 2": true,
              "__name__ 3": true,
              "cluster_id 1": true,
              "cluster_id 2": true,
              "cluster_id 3": true,
              "cluster_name 2": true,
              "cluster_name 3": true,
              "exported_instance 1": true,
              "exported_instance 2": true,
              "exported_instance 3": true,
              "gpu_chip_index 1": false,
              "gpu_chip_index 2": true,
              "gpu_chip_index 3": true,
              "gpu_index 2": true,
              "gpu_index 3": true,
              "gpu_name 2": true,
              "gpu_name 3": true,
              "instance 1": true,
              "instance 2": true,
              "instance 3": true,
              "job 1": true,
              "job 2": true,
              "job 3": true,
              "worker_id 1": true,
              "worker_id 2": true,
              "worker_id 3": true,
              "worker_name 2": true,
              "worker_name 3": true
            },
            "includeByName": {},
            "indexByName": {
              "Time": 15,
              "Value #A": 12,
              "Value #B": 27,
              "Value #C": 11,
              "__name__ 1": 0,
              "__name__ 2": 16,
              "__name__ 3": 28,
              "cluster_id 1": 1,
              "cluster_id 2": 17,
              "cluster_id 3": 29,
              "cluster_name 1": 2,
              "cluster_name 2": 18,
              "cluster_name 3": 30,
              "exported_instance 1": 4,
              "exported_instance 2": 19,
              "exported_instance 3": 31,
              "gpu_chip_index 1": 5,
              "gpu_chip_index 2": 20,
              "gpu_chip_index 3": 32,
              "gpu_index 1": 6,
              "gpu_index 2": 21,
              "gpu_index 3": 33,
              "gpu_name 1": 7,
              "gpu_name 2": 22,
              "gpu_name 3": 34,
              "instance 1": 8,
              "instance 2": 23,
              "instance 3": 35,
              "job 1": 9,
              "job 2": 24,
              "job 3": 36,
              "worker_id 1": 10,
              "worker_id 2": 25,
              "worker_id 3": 37,
              "worker_name 1": 3,
              "worker_name 2": 26,
              "worker_name 3": 38
            },
            "renameByName": {
              "Value #A": "GPU Utilization",
              "Value #B": "GPU VRAM Utilization",
              "Value #C": "Temperature",
              "cluster_name 1": "Cluster",
              "exported_instance 1": "",
              "gpu_chip_index 1": "Chip Index",
              "gpu_index 1": "GPU Index",
              "gpu_name 1": "GPU Name",
              "worker_name 1": "Worker"
            }
          }
        }
      ],
      "type": "table"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "custom": {
            "axisBorderShow": false,
            "axisCenteredZero": false,
            "axisColorMode": "text",
            "axisLabel": "",
            "axisPlacement": "auto",
            "barAlignment": 0,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 10,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineWidth": 2,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": true,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "mappings": [],
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "bytes"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 6,
        "x": 0,
        "y": 15
      },
      "id": 92,
      "options": {
        "dataLinks": [],
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "multi",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_memory_used_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "hide": false,
          "instant": false,
          "interval": "",
          "intervalFactor": 2,
          "legendFormat": "{{worker_name}}",
          "refId": "A"
        }
      ],
      "title": "System Memory Usage",
      "type": "timeseries"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "description": "",
      "fieldConfig": {
        "defaults": {
          "color": {
            "fixedColor": "blue",
            "mode": "palette-classic"
          },
          "custom": {
            "axisBorderShow": false,
            "axisCenteredZero": false,
            "axisColorMode": "text",
            "axisLabel": "",
            "axisPlacement": "auto",
            "barAlignment": 0,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 10,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineWidth": 1,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": false,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "fieldMinMax": false,
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 9,
        "x": 6,
        "y": 15
      },
      "id": 48,
      "options": {
        "legend": {
          "calcs": [
            "max",
            "last"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "gpustack:worker_node_gpu_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "instant": false,
          "legendFormat": "{{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "GPU Utilization",
      "type": "timeseries"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "description": "",
      "fieldConfig": {
        "defaults": {
          "color": {
            "fixedColor": "blue",
            "mode": "palette-classic"
          },
          "custom": {
            "axisBorderShow": false,
            "axisCenteredZero": false,
            "axisColorMode": "text",
            "axisLabel": "",
            "axisPlacement": "auto",
            "barAlignment": 0,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 10,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineStyle": {
              "fill": "solid"
            },
            "lineWidth": 1,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": false,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "fieldMinMax": false,
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 9,
        "x": 15,
        "y": 15
      },
      "id": 86,
      "options": {
        "legend": {
          "calcs": [
            "max",
            "last"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "gpustack:worker_node_gram_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "instant": false,
          "legendFormat": "{{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "VRAM 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,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 10,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineWidth": 2,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": false,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 6,
        "x": 0,
        "y": 28
      },
      "id": 91,
      "options": {
        "dataLinks": [],
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "multi",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_cpu_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "intervalFactor": 1,
          "legendFormat": "{{worker_name}}",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "CPU 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,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 10,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineWidth": 1,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": true,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "fieldMinMax": false,
          "mappings": [],
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "bytes"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 9,
        "x": 6,
        "y": 28
      },
      "id": 88,
      "options": {
        "dataLinks": [],
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "multi",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_gram_total_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "hide": false,
          "intervalFactor": 1,
          "legendFormat": "Total: {{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        },
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_gram_used_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "hide": false,
          "instant": false,
          "legendFormat": "Used: {{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "B"
        },
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_gram_allocated_bytes{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "hide": false,
          "instant": false,
          "legendFormat": "Allocated: {{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "C"
        }
      ],
      "title": "GPU Mem Usage",
      "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,
            "barWidthFactor": 0.6,
            "drawStyle": "line",
            "fillOpacity": 0,
            "gradientMode": "none",
            "hideFrom": {
              "legend": false,
              "tooltip": false,
              "viz": false
            },
            "insertNulls": false,
            "lineInterpolation": "linear",
            "lineWidth": 2,
            "pointSize": 5,
            "scaleDistribution": {
              "type": "linear"
            },
            "showPoints": "never",
            "showValues": false,
            "spanNulls": true,
            "stacking": {
              "group": "A",
              "mode": "none"
            },
            "thresholdsStyle": {
              "mode": "off"
            }
          },
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "celsius"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 13,
        "w": 9,
        "x": 15,
        "y": 28
      },
      "id": 87,
      "options": {
        "dataLinks": [],
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "hideZeros": false,
          "mode": "multi",
          "sort": "none"
        }
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "expr": "gpustack:worker_node_gpu_temperature_celsius{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}",
          "format": "time_series",
          "hide": false,
          "intervalFactor": 1,
          "legendFormat": "{{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "GPU Temperature",
      "type": "timeseries"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "decimals": 2,
          "mappings": [],
          "max": 100,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 14,
        "w": 12,
        "x": 0,
        "y": 41
      },
      "id": 78,
      "options": {
        "displayMode": "lcd",
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "maxVizHeight": 300,
        "minVizHeight": 16,
        "minVizWidth": 8,
        "namePlacement": "auto",
        "orientation": "auto",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "showUnfilled": true,
        "sizing": "auto",
        "valueMode": "color"
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          },
          "editorMode": "code",
          "exemplar": false,
          "expr": "avg by (cluster_name,worker_name,gpu_chip_index,gpu_index) (gpustack:worker_node_gpu_utilization_rate{cluster_name=~\"$cluster_name\",worker_name=~\"$worker_name\"}) ",
          "instant": false,
          "legendFormat": "{{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "GPU Utilization",
      "type": "bargauge"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "decimals": 2,
          "mappings": [],
          "max": 100,
          "min": 0,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "green",
                "value": 0
              },
              {
                "color": "#EAB839",
                "value": 50
              },
              {
                "color": "#EF843C",
                "value": 60
              },
              {
                "color": "red",
                "value": 80
              }
            ]
          },
          "unit": "percent"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 14,
        "w": 12,
        "x": 12,
        "y": 41
      },
      "id": 79,
      "options": {
        "displayMode": "gradient",
        "legend": {
          "calcs": [
            "last",
            "max"
          ],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "maxVizHeight": 300,
        "minVizHeight": 16,
        "minVizWidth": 8,
        "namePlacement": "auto",
        "orientation": "auto",
        "reduceOptions": {
          "calcs": [
            "mean"
          ],
          "fields": "",
          "values": false
        },
        "showUnfilled": true,
        "sizing": "auto",
        "valueMode": "color"
      },
      "pluginVersion": "12.2.0",
      "targets": [
        {
          "datasource": {
            "type": "prometheus",
            "uid": "${idc}"
          },
          "editorMode": "code",
          "expr": "avg by (cluster_name,worker_name, gpu_chip_index,gpu_index) (gpustack:worker_node_gram_utilization_rate{cluster_name=\"$cluster_name\",worker_name=~\"$worker_name\"}) ",
          "instant": false,
          "legendFormat": "{{worker_name}} GPU{{gpu_index}} (chip {{gpu_chip_index}})",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "VRAM Utilization",
      "type": "bargauge"
    }
  ],
  "preload": false,
  "refresh": "30s",
  "schemaVersion": 42,
  "tags": [],
  "templating": {
    "list": [
      {
        "current": {
          "text": "",
          "value": ""
        },
        "datasource": {
          "type": "prometheus",
          "uid": "prometheus"
        },
        "definition": "query_result(gpustack:cluster_info)",
        "includeAll": false,
        "label": "Cluster",
        "name": "cluster_name",
        "options": [],
        "query": {
          "qryType": 1,
          "query": "query_result(gpustack:cluster_info)",
          "refId": "PrometheusVariableQueryEditor-VariableQuery"
        },
        "refresh": 2,
        "regex": "/cluster_name=\"([^\"]+)\"/",
        "sort": 1,
        "type": "query"
      },
      {
        "allValue": ".*",
        "current": {
          "text": "All",
          "value": [
            "$__all"
          ]
        },
        "datasource": {
          "type": "prometheus",
          "uid": "prometheus"
        },
        "definition": "query_result(gpustack:worker_info{cluster_name=\"$cluster_name\"})",
        "includeAll": true,
        "label": "Worker",
        "multi": true,
        "name": "worker_name",
        "options": [],
        "query": {
          "qryType": 1,
          "query": "query_result(gpustack:worker_info{cluster_name=\"$cluster_name\"})",
          "refId": "PrometheusVariableQueryEditor-VariableQuery"
        },
        "refresh": 2,
        "regex": "/worker_name=\"([^\"]+)\"/",
        "type": "query"
      }
    ]
  },
  "time": {
    "from": "now-5m",
    "to": "now"
  },
  "timepicker": {
    "refresh_intervals": [
      "10s",
      "30s",
      "1m",
      "5m",
      "15m",
      "30m",
      "1h",
      "2h"
    ]
  },
  "timezone": "browser",
  "title": "GPUStack Worker",
  "uid": "gpustack-worker",
  "version": 1
}