groups:
  - name: system
    rules:
      - alert: HostHighCPU
#        expr: rate(node_cpu_seconds_total{mode!="idle"}[2m]) > 0.9
        expr: 100 * (1 - avg by(instance) (rate(node_cpu_seconds_total{mode="idle"}[2m]))) > 90
        for: 2m
        labels:
          severity: warning
        annotations:
          summary: "High CPU usage on host"

      - alert: ContainerRestarting
#        expr: increase(container_start_time_seconds[10m]) > 3
        expr: increase(container_restart_count[10m]) > 3
        for: 1m
        labels:
          severity: warning
        annotations:
          summary: "Container restarting frequently"

      - alert: AlwaysFiring
        expr: vector(1)
        for: 10s
        labels:
          severity: critical
        annotations:
          summary: "This alert should always fire"