{
  "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": null,
  "links": [],
  "liveNow": true,
  "panels": [
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "thresholds"
          },
          "mappings": [],
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {
                "color": "red",
                "value": null
              },
              {
                "color": "green",
                "value": 1
              }
            ]
          },
          "unit": "short"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 6,
        "w": 24,
        "x": 0,
        "y": 0
      },
      "id": 1,
      "options": {
        "colorMode": "background",
        "graphMode": "none",
        "justifyMode": "auto",
        "orientation": "auto",
        "reduceOptions": {
          "calcs": [
            "lastNotNull"
          ],
          "fields": "",
          "values": false
        },
        "textMode": "auto"
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "up{job=~\"prometheus|llama_exporter\"}",
          "instant": true,
          "legendFormat": "{{job}}",
          "range": false,
          "refId": "A"
        }
      ],
      "title": "LLM Exporter Health",
      "type": "stat"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "short"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 0,
        "y": 6
      },
      "id": 2,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "llama_server_health{model=~\"$model\"}",
          "legendFormat": "health {{server}}",
          "range": true,
          "refId": "A"
        },
        {
          "editorMode": "code",
          "expr": "llama_models_loaded{model=~\"$model\"}",
          "legendFormat": "models {{server}} {{model}}",
          "range": true,
          "refId": "B"
        }
      ],
      "title": "Llama Server Metrics",
      "type": "timeseries"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "reqps"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 12,
        "y": 6
      },
      "id": 3,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "rate(llamacpp:tokens_predicted_total{model=~\"$model\"}[5m])",
          "legendFormat": "decode {{model}}",
          "range": true,
          "refId": "A"
        },
        {
          "editorMode": "code",
          "expr": "rate(llamacpp:prompt_tokens_total{model=~\"$model\"}[5m])",
          "legendFormat": "prompt {{model}}",
          "range": true,
          "refId": "B"
        }
      ],
      "title": "Token Throughput",
      "type": "timeseries"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "s"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 0,
        "y": 14
      },
      "id": 4,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "rate(llamacpp:tokens_predicted_seconds_total{model=~\"$model\"}[5m])",
          "legendFormat": "decode time {{model}}",
          "range": true,
          "refId": "A"
        },
        {
          "editorMode": "code",
          "expr": "rate(llamacpp:prompt_seconds_total{model=~\"$model\"}[5m])",
          "legendFormat": "prompt time {{model}}",
          "range": true,
          "refId": "B"
        }
      ],
      "title": "Processing Time",
      "type": "timeseries"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "short"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 12,
        "y": 14
      },
      "id": 5,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "llamacpp:requests_processing{model=~\"$model\"}",
          "legendFormat": "processing {{model}}",
          "range": true,
          "refId": "A"
        },
        {
          "editorMode": "code",
          "expr": "llamacpp:requests_deferred{model=~\"$model\"}",
          "legendFormat": "deferred {{model}}",
          "range": true,
          "refId": "B"
        }
      ],
      "title": "Request Queue",
      "type": "timeseries"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "short"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 0,
        "y": 22
      },
      "id": 6,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "llamacpp:n_busy_slots_per_decode{model=~\"$model\"}",
          "legendFormat": "busy slots {{model}}",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Busy Slots Per Decode",
      "type": "timeseries"
    },
    {
      "datasource": "$datasource",
      "fieldConfig": {
        "defaults": {
          "color": {
            "mode": "palette-classic"
          },
          "unit": "short"
        },
        "overrides": []
      },
      "gridPos": {
        "h": 8,
        "w": 12,
        "x": 12,
        "y": 22
      },
      "id": 7,
      "options": {
        "legend": {
          "calcs": [],
          "displayMode": "table",
          "placement": "bottom",
          "showLegend": true
        },
        "tooltip": {
          "mode": "single",
          "sort": "none"
        }
      },
      "pluginVersion": "11.0.0",
      "targets": [
        {
          "editorMode": "code",
          "expr": "llamacpp:n_tokens_max{model=~\"$model\"}",
          "legendFormat": "max tokens {{model}}",
          "range": true,
          "refId": "A"
        }
      ],
      "title": "Max Tokens (Context Window)",
      "type": "timeseries"
    }
  ],
  "refresh": "5s",
  "schemaVersion": 39,
  "style": "dark",
  "tags": [
    "llm",
    "monitoring",
    "realtime"
  ],
  "templating": {
    "list": [
      {
        "current": {
          "selected": true,
          "text": "Prometheus",
          "value": "Prometheus"
        },
        "hide": 0,
        "includeAll": false,
        "label": "Datasource",
        "multi": false,
        "name": "datasource",
        "options": [],
        "query": "prometheus",
        "refresh": 1,
        "regex": "",
        "skipUrlSync": false,
        "type": "datasource"
      },
      {
        "current": {
          "selected": false,
          "text": "All",
          "value": "$__all"
        },
        "datasource": {
          "type": "prometheus",
          "uid": "$datasource"
        },
        "definition": "label_values(llamacpp:tokens_predicted_total, model)",
        "hide": 0,
        "includeAll": true,
        "label": "Model",
        "multi": false,
        "name": "model",
        "options": [],
        "query": "label_values(llamacpp:tokens_predicted_total, model)",
        "refresh": 2,
        "regex": "",
        "skipUrlSync": false,
        "sort": 1,
        "type": "query"
      }
    ]
  },
  "time": {
    "from": "now-30m",
    "to": "now"
  },
  "timepicker": {},
  "timezone": "",
  "title": "LLM Monitoring Realtime",
  "uid": "llm-monitoring-realtime",
  "version": 1,
  "weekStart": ""
}