groups: - name: pg_local_cache.availability rules: - alert: PgLocalCacheExporterDown expr: up{job="pg-local-cache"} == 0 for: 2m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache exporter is unavailable description: Prometheus has been unable to scrape pg_local_cache for two minutes. - alert: PgLocalCacheMetricsMissing expr: up{job="pg-local-cache"} == 1 unless on (instance) pg_local_cache_up{job="pg-local-cache"} for: 2m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache custom metrics are missing description: postgres_exporter is reachable but local_cache.metrics() is not being exported. - alert: PgLocalCacheUnavailable expr: pg_local_cache_up{job="pg-local-cache"} == 0 for: 1m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache reports unavailable description: The extension metrics endpoint reports that shared state is not ready. - alert: PgLocalCacheWorkersMissing expr: pg_local_cache_workers_running < pg_local_cache_workers_configured for: 2m labels: severity: critical component: pg_local_cache annotations: summary: One or more pg_local_cache workers are missing description: The number of running RESP workers is below the configured worker count. - alert: PgLocalCacheMappingsIncomplete expr: pg_local_cache_workers_with_incomplete_mappings > 0 for: 1m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache mappings are not current on every worker description: At least one RESP worker has not loaded the current complete table-mapping generation. - name: pg_local_cache.capacity rules: - alert: PgLocalCacheMemoryBudgetHigh expr: pg_local_cache_estimated_memory_bytes / clamp_min(pg_local_cache_memory_budget_bytes, 1) > 0.85 for: 30m labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache memory budget has little headroom description: The deterministic extension memory estimate exceeds 85% of its startup budget. - alert: PgLocalCacheMemoryBudgetCritical expr: pg_local_cache_estimated_memory_bytes / clamp_min(pg_local_cache_memory_budget_bytes, 1) > 0.95 for: 5m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache memory budget is almost exhausted description: The deterministic extension memory estimate exceeds 95% of its startup budget. - alert: PgLocalCacheNearCapacity expr: pg_local_cache_entries / clamp_min(pg_local_cache_cache_capacity, 1) > 0.90 for: 15m labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache is near entry capacity description: More than 90% of cache entry slots have been occupied for 15 minutes. - alert: PgLocalCacheAtCapacity expr: pg_local_cache_entries / clamp_min(pg_local_cache_cache_capacity, 1) > 0.99 for: 5m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache is at entry capacity description: More than 99% of cache entry slots have been occupied for five minutes. - alert: PgLocalCacheClientSaturation expr: pg_local_cache_active_clients / clamp_min(pg_local_cache_max_clients, 1) > 0.80 for: 5m labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache client capacity is under pressure description: Active RESP clients have exceeded 80% of the global limit for five minutes. - alert: PgLocalCacheClientSaturationCritical expr: pg_local_cache_active_clients / clamp_min(pg_local_cache_max_clients, 1) > 0.95 for: 2m labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache client capacity is almost exhausted description: Active RESP clients have exceeded 95% of the global limit for two minutes. - name: pg_local_cache.failures rules: - alert: PgLocalCacheClientLimitRejections expr: increase(pg_local_cache_client_limit_rejections_total[5m]) > 0 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache rejected clients at the global limit description: At least one RESP connection was rejected by the OOM safety limit in five minutes. - alert: PgLocalCacheConnectionRejectionsHigh expr: increase(pg_local_cache_rejected_connections_total[5m]) > 10 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache connection rejection rate is high description: More than ten RESP connections were rejected in five minutes. - alert: PgLocalCacheAuthenticationFailures expr: increase(pg_local_cache_authentication_failures_total[5m]) > 5 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache authentication failures detected description: More than five RESP authentication attempts failed in five minutes. - alert: PgLocalCacheSlowClientDrops expr: increase(pg_local_cache_slow_client_drops_total[5m]) > 0 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache dropped slow clients description: At least one slow RESP client was disconnected in five minutes. - alert: PgLocalCacheDirtyKeyFallback expr: increase(pg_local_cache_dirty_key_limit_fallbacks_total[10m]) > 0 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache transaction dirty-key limit was reached description: At least one transaction used relation-wide invalidation after reaching its key limit. - alert: PgLocalCacheMappingReloadFailure expr: increase(pg_local_cache_mapping_reload_failures_total[5m]) > 0 labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache mapping reload failed description: A mapping reload failure occurred in the last five minutes. - alert: PgLocalCacheMappingReloadIncomplete expr: increase(pg_local_cache_mapping_reload_incomplete_retries_total[5m]) > 0 labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache retried an incomplete mapping set description: At least one worker omitted an unsafe mapping during reload in the last five minutes. - name: pg_local_cache.container_oom rules: - alert: PgLocalCacheContainerMemoryPressure expr: | container_memory_working_set_bytes{name=~".*pg_local_cache.*postgres.*"} / clamp_min(container_spec_memory_limit_bytes{name=~".*pg_local_cache.*postgres.*"}, 1) > 0.90 for: 10m labels: severity: warning component: pg_local_cache annotations: summary: pg_local_cache PostgreSQL container memory pressure description: The PostgreSQL container has used more than 90% of its memory limit for ten minutes. - alert: PgLocalCacheContainerOOMKilled expr: increase(container_oom_events_total{name=~".*pg_local_cache.*postgres.*"}[10m]) > 0 labels: severity: critical component: pg_local_cache annotations: summary: pg_local_cache PostgreSQL container had an OOM event description: cAdvisor reported an OOM event for the PostgreSQL container in the last ten minutes.