food-market/deploy/grafana/dashboards/quality-watchdog.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"
      }
    ]
  },
  "description": "Sprint 26: quality-watchdog dashboard. Метрики из ~/quality-watchdog.sh (textfile exporter, см. docs/observability.md) + базовые food-market.api метрики (/metrics).",
  "editable": true,
  "fiscalYearStartMonth": 0,
  "graphTooltip": 1,
  "id": null,
  "links": [],
  "liveNow": false,
  "panels": [
    {
      "id": 1,
      "type": "stat",
      "title": "Smoke success ratio (7 дней)",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 5, "w": 6, "x": 0, "y": 0},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "thresholds"},
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "red", "value": null},
              {"color": "orange", "value": 0.80},
              {"color": "green", "value": 0.95}
            ]
          },
          "unit": "percentunit",
          "min": 0,
          "max": 1
        }
      },
      "options": {
        "graphMode": "area",
        "colorMode": "value",
        "justifyMode": "center",
        "reduceOptions": {"calcs": ["lastNotNull"]}
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum(increase(quality_watchdog_run_total{result=\"green\"}[7d])) / sum(increase(quality_watchdog_run_total[7d]))",
          "legendFormat": "green ratio"
        }
      ]
    },
    {
      "id": 2,
      "type": "stat",
      "title": "Incidents (7 дней)",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 5, "w": 6, "x": 6, "y": 0},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "thresholds"},
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "green", "value": null},
              {"color": "orange", "value": 1},
              {"color": "red", "value": 3}
            ]
          },
          "unit": "short"
        }
      },
      "options": {"colorMode": "value", "reduceOptions": {"calcs": ["lastNotNull"]}},
      "targets": [
        {
          "refId": "A",
          "expr": "sum(increase(quality_watchdog_incidents_total[7d]))",
          "legendFormat": "incidents"
        }
      ]
    },
    {
      "id": 3,
      "type": "stat",
      "title": "Multi-tenant violations (24h) — должно быть 0",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 5, "w": 6, "x": 12, "y": 0},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "thresholds"},
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "green", "value": null},
              {"color": "red", "value": 1}
            ]
          },
          "unit": "short"
        }
      },
      "options": {"colorMode": "value", "reduceOptions": {"calcs": ["lastNotNull"]}},
      "targets": [
        {
          "refId": "A",
          "expr": "sum(increase(quality_watchdog_step_failure_total{step=\"multi_tenant\"}[24h]))",
          "legendFormat": "leaks"
        }
      ]
    },
    {
      "id": 4,
      "type": "stat",
      "title": "Текущий статус watchdog",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 5, "w": 6, "x": 18, "y": 0},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "thresholds"},
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "red", "value": null},
              {"color": "orange", "value": 0.5},
              {"color": "green", "value": 1}
            ]
          },
          "unit": "short",
          "mappings": [
            {"options": {"0": {"text": "🔴 RED"}, "1": {"text": "🟢 GREEN"}}, "type": "value"}
          ]
        }
      },
      "options": {"colorMode": "background", "reduceOptions": {"calcs": ["lastNotNull"]}},
      "targets": [
        {
          "refId": "A",
          "expr": "quality_watchdog_last_run_status",
          "legendFormat": "status"
        }
      ]
    },
    {
      "id": 5,
      "type": "timeseries",
      "title": "p95 latency по endpoint (мс)",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 9, "w": 12, "x": 0, "y": 5},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 10,
            "showPoints": "never",
            "spanNulls": false,
            "stacking": {"mode": "none"}
          },
          "unit": "ms",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "green", "value": null},
              {"color": "orange", "value": 400},
              {"color": "red", "value": 800}
            ]
          }
        }
      },
      "options": {"legend": {"displayMode": "table", "placement": "bottom"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "quality_watchdog_endpoint_p95_ms",
          "legendFormat": "{{endpoint}}"
        }
      ]
    },
    {
      "id": 6,
      "type": "timeseries",
      "title": "Шаги watchdog — pass/fail во времени",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 9, "w": 12, "x": 12, "y": 5},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {
            "drawStyle": "bars",
            "lineWidth": 1,
            "fillOpacity": 60,
            "stacking": {"mode": "normal"}
          },
          "unit": "short"
        }
      },
      "options": {"legend": {"displayMode": "list", "placement": "right"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "sum by (step) (increase(quality_watchdog_step_failure_total[1h]))",
          "legendFormat": "{{step}}"
        }
      ]
    },
    {
      "id": 7,
      "type": "timeseries",
      "title": "HTTP request rate (rps)",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 8, "w": 12, "x": 0, "y": 14},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {"drawStyle": "line", "lineWidth": 2, "fillOpacity": 10},
          "unit": "reqps"
        }
      },
      "options": {"legend": {"displayMode": "list"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "sum(rate(http_requests_received_total[5m])) by (code)",
          "legendFormat": "code={{code}}"
        }
      ]
    },
    {
      "id": 8,
      "type": "timeseries",
      "title": "DB query duration p95 (food_market_db_query_duration_seconds)",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 8, "w": 12, "x": 12, "y": 14},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {"drawStyle": "line", "lineWidth": 2, "fillOpacity": 10},
          "unit": "s",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "green", "value": null},
              {"color": "orange", "value": 0.5},
              {"color": "red", "value": 1.0}
            ]
          }
        }
      },
      "options": {"legend": {"displayMode": "list"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "histogram_quantile(0.95, sum(rate(food_market_db_query_duration_seconds_bucket[5m])) by (le))",
          "legendFormat": "p95 DB"
        }
      ]
    },
    {
      "id": 9,
      "type": "timeseries",
      "title": "Документы проведены / ошибки",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 8, "w": 12, "x": 0, "y": 22},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {"drawStyle": "line", "lineWidth": 2, "fillOpacity": 10},
          "unit": "ops"
        }
      },
      "options": {"legend": {"displayMode": "table", "placement": "bottom"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "sum(rate(food_market_documents_posted_total[5m])) by (type)",
          "legendFormat": "posted {{type}}"
        },
        {
          "refId": "B",
          "expr": "sum(rate(food_market_documents_error_total[5m])) by (type)",
          "legendFormat": "error {{type}}"
        }
      ]
    },
    {
      "id": 10,
      "type": "timeseries",
      "title": "Свободное место на диске",
      "datasource": {"type": "prometheus", "uid": "${DS_PROMETHEUS}"},
      "gridPos": {"h": 8, "w": 12, "x": 12, "y": 22},
      "fieldConfig": {
        "defaults": {
          "color": {"mode": "palette-classic"},
          "custom": {"drawStyle": "line", "lineWidth": 2, "fillOpacity": 10},
          "unit": "bytes",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              {"color": "red", "value": null},
              {"color": "orange", "value": 5368709120},
              {"color": "green", "value": 10737418240}
            ]
          }
        }
      },
      "options": {"legend": {"displayMode": "list"}, "tooltip": {"mode": "multi"}},
      "targets": [
        {
          "refId": "A",
          "expr": "food_market_disk_free_bytes",
          "legendFormat": "{{mount}}"
        }
      ]
    }
  ],
  "refresh": "1m",
  "schemaVersion": 38,
  "tags": ["food-market", "quality-watchdog", "sprint26"],
  "templating": {
    "list": [
      {
        "current": {"selected": false, "text": "Prometheus", "value": "Prometheus"},
        "hide": 0,
        "includeAll": false,
        "label": "Datasource",
        "multi": false,
        "name": "DS_PROMETHEUS",
        "options": [],
        "query": "prometheus",
        "queryValue": "",
        "refresh": 1,
        "regex": "",
        "skipUrlSync": false,
        "type": "datasource"
      }
    ]
  },
  "time": {"from": "now-7d", "to": "now"},
  "timepicker": {},
  "timezone": "browser",
  "title": "food-market — quality-watchdog",
  "uid": "fm-quality-watchdog",
  "version": 1
}