11.6Monitoring Error Budget Consumption

Prometheus/Grafana Setup

# prometheus-error-budget-recording-rules.yaml
groups:
  - name: error-budget
    interval: 1m
    rules:
      # SLI: availability (successful requests / total requests)
      - record: sli:checkout:availability:5m
        expr: |
          sum(rate(http_requests_total{service="checkout",status!~"5.."}[5m]))
          /
          sum(rate(http_requests_total{service="checkout"}[5m]))

      # Error budget remaining (30-day window)
      - record: error_budget:checkout:availability:remaining
        expr: |
          1 - (
            (1 - sli:checkout:availability:30d) / (1 - 0.9995)
          )
        # Result: 1.0 = full budget, 0.0 = budget exhausted, <0 = over budget

      # Error budget burn rate
      - record: error_budget:checkout:availability:burn_rate:1h
        expr: |
          (1 - sli:checkout:availability:1h) / (1 - 0.9995)
        # Result: 1.0 = sustainable rate, >1.0 = burning faster than allowed