veylant/deploy/grafana/dashboards/production-slo.json

{
  "title": "Veylant — Production SLO & Error Budget",
  "uid": "veylant-production-slo",
  "schemaVersion": 38,
  "version": 1,
  "refresh": "1m",
  "time": { "from": "now-30d", "to": "now" },
  "tags": ["slo", "production", "veylant"],
  "panels": [
    {
      "id": 1,
      "title": "Uptime SLO — 30-day rolling (target: 99.5%)",
      "type": "gauge",
      "gridPos": { "h": 8, "w": 6, "x": 0, "y": 0 },
      "options": {
        "reduceOptions": { "calcs": ["lastNotNull"] },
        "orientation": "auto",
        "showThresholdLabels": true,
        "showThresholdMarkers": true
      },
      "fieldConfig": {
        "defaults": {
          "unit": "percentunit",
          "min": 0.99,
          "max": 1,
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "red", "value": null },
              { "color": "yellow", "value": 0.995 },
              { "color": "green", "value": 0.999 }
            ]
          }
        }
      },
      "targets": [
        {
          "expr": "1 - (sum(increase(veylant_request_errors_total[30d])) / sum(increase(veylant_requests_total[30d])))",
          "legendFormat": "Uptime SLO"
        }
      ]
    },
    {
      "id": 2,
      "title": "Error Budget Remaining (minutes)",
      "description": "SLO target: 99.5% uptime over 30 days = 216 min allowed downtime",
      "type": "stat",
      "gridPos": { "h": 8, "w": 6, "x": 6, "y": 0 },
      "options": {
        "reduceOptions": { "calcs": ["lastNotNull"] },
        "colorMode": "background"
      },
      "fieldConfig": {
        "defaults": {
          "unit": "m",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "red", "value": null },
              { "color": "yellow", "value": 43 },
              { "color": "green", "value": 108 }
            ]
          }
        }
      },
      "targets": [
        {
          "expr": "(0.005 * 30 * 24 * 60) - (sum(increase(veylant_request_errors_total[30d])) / sum(increase(veylant_requests_total[30d])) * 30 * 24 * 60)",
          "legendFormat": "Budget remaining (min)"
        }
      ]
    },
    {
      "id": 3,
      "title": "p99 Latency SLO (target: < 500ms)",
      "type": "gauge",
      "gridPos": { "h": 8, "w": 6, "x": 12, "y": 0 },
      "options": {
        "reduceOptions": { "calcs": ["lastNotNull"] },
        "orientation": "auto",
        "showThresholdMarkers": true
      },
      "fieldConfig": {
        "defaults": {
          "unit": "s",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "green", "value": null },
              { "color": "yellow", "value": 0.3 },
              { "color": "red", "value": 0.5 }
            ]
          }
        }
      },
      "targets": [
        {
          "expr": "histogram_quantile(0.99, sum by (le) (rate(veylant_request_duration_seconds_bucket[5m])))",
          "legendFormat": "p99 latency"
        }
      ]
    },
    {
      "id": 4,
      "title": "Active Alerts",
      "type": "stat",
      "gridPos": { "h": 8, "w": 6, "x": 18, "y": 0 },
      "options": {
        "reduceOptions": { "calcs": ["lastNotNull"] },
        "colorMode": "background"
      },
      "fieldConfig": {
        "defaults": {
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "green", "value": null },
              { "color": "yellow", "value": 1 },
              { "color": "red", "value": 3 }
            ]
          }
        }
      },
      "targets": [
        {
          "expr": "sum(ALERTS{alertstate=\"firing\",job=~\"veylant.*\"})",
          "legendFormat": "Firing alerts"
        }
      ]
    },
    {
      "id": 5,
      "title": "PII Entities Detected — Rate by Type (per min)",
      "type": "timeseries",
      "gridPos": { "h": 8, "w": 12, "x": 0, "y": 8 },
      "targets": [
        {
          "expr": "sum by (entity_type) (rate(veylant_pii_entities_detected_total[1m])) * 60",
          "legendFormat": "{{ entity_type }}"
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "short",
          "custom": { "lineWidth": 2 }
        }
      }
    },
    {
      "id": 6,
      "title": "PostgreSQL Active Connections",
      "type": "timeseries",
      "gridPos": { "h": 8, "w": 12, "x": 12, "y": 8 },
      "targets": [
        {
          "expr": "veylant_db_connections_active",
          "legendFormat": "Active connections"
        },
        {
          "expr": "veylant_db_connections_idle",
          "legendFormat": "Idle connections"
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "short",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "green", "value": null },
              { "color": "yellow", "value": 15 },
              { "color": "red", "value": 20 }
            ]
          }
        }
      }
    },
    {
      "id": 7,
      "title": "Provider RPS Breakdown",
      "type": "piechart",
      "gridPos": { "h": 8, "w": 8, "x": 0, "y": 16 },
      "options": {
        "pieType": "donut",
        "displayLabels": ["name", "percent"]
      },
      "targets": [
        {
          "expr": "sum by (provider) (rate(veylant_requests_total[5m]))",
          "legendFormat": "{{ provider }}"
        }
      ]
    },
    {
      "id": 8,
      "title": "Provider RPS — Time Series",
      "type": "timeseries",
      "gridPos": { "h": 8, "w": 16, "x": 8, "y": 16 },
      "targets": [
        {
          "expr": "sum by (provider) (rate(veylant_requests_total[1m]))",
          "legendFormat": "{{ provider }}"
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "reqps",
          "custom": { "lineWidth": 2 }
        }
      }
    },
    {
      "id": 9,
      "title": "Redis Memory Usage %",
      "type": "timeseries",
      "gridPos": { "h": 8, "w": 12, "x": 0, "y": 24 },
      "targets": [
        {
          "expr": "redis_memory_used_bytes / redis_memory_max_bytes * 100",
          "legendFormat": "Redis memory %"
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "percent",
          "thresholds": {
            "mode": "absolute",
            "steps": [
              { "color": "green", "value": null },
              { "color": "yellow", "value": 70 },
              { "color": "red", "value": 90 }
            ]
          }
        }
      }
    },
    {
      "id": 10,
      "title": "Error Rate by Provider (5m avg)",
      "type": "timeseries",
      "gridPos": { "h": 8, "w": 12, "x": 12, "y": 24 },
      "targets": [
        {
          "expr": "veylant:error_rate:5m * 100",
          "legendFormat": "{{ provider }} error %"
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "percent",
          "custom": { "lineWidth": 2 }
        }
      }
    }
  ]
}