rules.yml

groups:
- name: replication
  rules:
  # docker-compose exec replica gosu postgres psql -c "SELECT pg_wal_replay_pause()"
  # docker-compose exec replica gosu postgres psql -c "SELECT pg_wal_replay_resume()"
  - alert: PGReplicationLagXLOG
    expr: max(pg_xlog_position) - min(pg_xlog_position) > 1000000000
    labels:
      severity: critical
    annotations:
      summary: Replication Lag
  - alert: PGReplicationLagTime
    expr: pg_replication_replag > 60000
    labels:
      severity: critical
    annotations:
      summary: Replication Lag more than 1 min
  - alert: PGReplicationLagBytes
    expr: pg_replication_replay_lag > 1000000
    labels:
      severity: critical
    annotations:
      summary: Replication Lag more than 10 MB
- name: connections
  rules:
  - alert: TooManyConnections
    expr: sum by (exported_server, instance, job, server) (pg_stat_database_numbackends) / pg_settings_max_connections > 0.9
    labels:
      severity: critical
    annotations:
      summary: Too many connections
- name: tables
  rules:
  - alert: PGFatIndex
    expr: pg_table_size_index_bytes{schemaname!~"pg_catalog|information_schema"} > (pg_table_size_table_bytes{schemaname!~"pg_catalog|information_schema"} * 10)
    labels:
      severity: critical
    annotations:
      summary: Index is far bigger than data, REINDEX needed