{
  "uid": "factory-ai",
  "title": "Factory — AI",
  "tags": ["factory", "observability", "ai"],
  "timezone": "browser",
  "schemaVersion": 39,
  "version": 1,
  "editable": true,
  "refresh": "30s",
  "time": { "from": "now-6h", "to": "now" },
  "panels": [
    {
      "id": 1,
      "type": "stat",
      "title": "Tokens / min (in, 5 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 4, "w": 6, "x": 0, "y": 0 },
      "fieldConfig": { "defaults": { "unit": "short", "thresholds": { "mode": "absolute", "steps": [ { "color": "green", "value": null } ] } }, "overrides": [] },
      "options": { "colorMode": "value", "graphMode": "area" },
      "targets": [ { "refId": "A", "expr": "sum by (model) (rate(ai_tokens_total{service_name=\"factory-api\", direction=\"in\"}[5m])) * 60", "legendFormat": "{{model}} in" } ]
    },
    {
      "id": 2,
      "type": "stat",
      "title": "Tokens / min (out, 5 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 4, "w": 6, "x": 6, "y": 0 },
      "fieldConfig": { "defaults": { "unit": "short", "thresholds": { "mode": "absolute", "steps": [ { "color": "blue", "value": null } ] } }, "overrides": [] },
      "options": { "colorMode": "value", "graphMode": "area" },
      "targets": [ { "refId": "A", "expr": "sum by (model) (rate(ai_tokens_total{service_name=\"factory-api\", direction=\"out\"}[5m])) * 60", "legendFormat": "{{model}} out" } ]
    },
    {
      "id": 3,
      "type": "timeseries",
      "title": "Tokens / min by direction + model",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 9, "w": 12, "x": 0, "y": 4 },
      "fieldConfig": { "defaults": { "unit": "short", "custom": { "drawStyle": "line", "fillOpacity": 10 } }, "overrides": [] },
      "options": { "legend": { "displayMode": "table", "placement": "right" } },
      "targets": [ { "refId": "A", "expr": "factory_api:ai_tokens:rate5m * 60", "legendFormat": "{{model}} {{direction}}" } ]
    },
    {
      "id": 4,
      "type": "timeseries",
      "title": "AI request p95 by model (10 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 9, "w": 12, "x": 12, "y": 4 },
      "fieldConfig": { "defaults": { "unit": "s", "custom": { "drawStyle": "line" }, "thresholds": { "mode": "absolute", "steps": [ { "color": "green", "value": null }, { "color": "red", "value": 5 } ] } }, "overrides": [] },
      "options": { "legend": { "displayMode": "list", "placement": "bottom" } },
      "targets": [ { "refId": "A", "expr": "factory_api:ai_request_duration_seconds:p95", "legendFormat": "{{model}}" } ]
    },
    {
      "id": 5,
      "type": "timeseries",
      "title": "AI request rate by model (5 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 8, "w": 8, "x": 0, "y": 13 },
      "fieldConfig": { "defaults": { "unit": "reqps", "custom": { "drawStyle": "line" } }, "overrides": [] },
      "options": { "legend": { "displayMode": "list", "placement": "bottom" } },
      "targets": [ { "refId": "A", "expr": "sum by (model) (rate(ai_request_duration_seconds_count{service_name=\"factory-api\"}[5m]))", "legendFormat": "{{model}}" } ]
    },
    {
      "id": 6,
      "type": "timeseries",
      "title": "AI request mean latency by model (5 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 8, "w": 8, "x": 8, "y": 13 },
      "fieldConfig": { "defaults": { "unit": "s", "custom": { "drawStyle": "line" } }, "overrides": [] },
      "options": { "legend": { "displayMode": "list", "placement": "bottom" } },
      "targets": [ { "refId": "A", "expr": "sum by (model) (rate(ai_request_duration_seconds_sum{service_name=\"factory-api\"}[5m])) / sum by (model) (rate(ai_request_duration_seconds_count{service_name=\"factory-api\"}[5m]))", "legendFormat": "{{model}}" } ]
    },
    {
      "id": 7,
      "type": "timeseries",
      "title": "AI errors by code (error_total, 5 m)",
      "datasource": { "type": "prometheus", "uid": "factory-prometheus" },
      "gridPos": { "h": 8, "w": 8, "x": 16, "y": 13 },
      "fieldConfig": { "defaults": { "unit": "ops", "custom": { "drawStyle": "bars" } }, "overrides": [] },
      "options": { "legend": { "displayMode": "list", "placement": "bottom" } },
      "targets": [ { "refId": "A", "expr": "sum by (code) (rate(error_total{service_name=\"factory-api\", code=~\"AI.*\"}[5m]))", "legendFormat": "{{code}}" } ]
    }
  ],
  "annotations": {
    "list": [
      {
        "builtIn": 1,
        "datasource": { "type": "grafana", "uid": "-- Grafana --" },
        "name": "Budget gauge — SPEC GAP",
        "description": "The spec's AI dashboard asks for a 'budget-approaching gauge from the usage rollup'. The ai:usage-rollup repeatable job writes budget usage to the DB (readable via the API) but exports NO Prometheus gauge — and budget is not in the spec's 14-metric list. Recorded in infrastructure/observability/README.md (spec-conflict, AGENTS rule 29); the token-rate panels above are the live proxy for budget burn.",
        "enable": true,
        "hide": false,
        "iconColor": "rgba(255, 96, 96, 1)"
      }
    ]
  }
}