// the find
samber/awesome-prometheus-alerts
🚨 Collection of Prometheus alerting rules
A community-maintained collection of 940+ copy-paste-ready Prometheus alerting rules across 90+ services — Kubernetes, MySQL, Redis, Kafka, and most things you'd find in a production stack. Aimed at SREs and platform engineers who want a starting point for alerting without writing PromQL from scratch. The Astro site makes rules browsable and copyable without touching the repo.
The breadth is the main value: coverage spans databases, message brokers, service meshes, cloud providers, and observability tooling itself (Thanos, Loki, OTel Collector). Rules are organized per-exporter, so if you're using the Percona MongoDB exporter you get exactly those rules, not a generic blob. There's a real build pipeline — source lives in `_data/rules.yml`, compiled output goes to `dist/` with CI tests, so the YAML you copy has actually been processed and validated rather than just hand-edited. Active maintenance with a commit as recent as June 2026 for a collection this size is genuinely unusual.
The thresholds are community opinions, not your numbers — a rule that fires at 'CPU > 80% for 5 minutes' will page you constantly on a batch workload and miss problems on a latency-sensitive API. There's no guidance on which thresholds to tune first, so you'll copy the YAML and then spend the first week silencing noisy alerts. Rules don't document their assumptions: some Kubernetes rules assume kube-state-metrics at a specific version, or specific label sets that may not match your cluster. When exporter metric names change between major versions, there's no changelog or deprecation notice per rule — you find out when your alert stops firing. No subscription mechanism for just the services you care about; tracking upstream changes to specific files is entirely on you.