monitoring-conventions · git:20261014.c8129da · 2026-10-14 · sha256 51df9f11a79051e6

monitoring-conventions git:20261014.c8129daA

Immutable. This exact content never changes and is served at /api/v1/blob/51df9f11a79051e6.

No tool to install first
claude plugin marketplace add https://markdownregistry.com/mp/a/art_oieyylva2qxvjzqo/51df9f11a79051e6bf4abed8eb926c603ce8fefa1cbe623273050dd3c3e348cf.json

Then /plugin install monitoring-conventions@markdownregistry-oieyylva2qxvjzqo in Claude Code. It installs this exact version, the one graded A here, pinned to the commit it was graded at, and leaves it there.

---
name: monitoring-conventions
description: Alert on symptoms, prevent alert fatigue, link runbooks, and separate health from readiness checks. Use when editing infrastructure code (CDK, Terraform, serverless config, CI workflows) or adding metrics and alerts.
---

# Monitoring and alerting rules

- Alert on SYMPTOMS such as error rate above 1%, p99 above 2 seconds, or a growing queue; do not page on causes such as CPU above 80% or memory use. Causes belong on dashboards.
- Prevent alert fatigue. If an alert fires three times without action, tune its threshold or remove it. An ignored alert is worse than no alert.
- Every P1 alert links to a runbook that says what to inspect, what to restart, and whom to escalate to. Do not add an alert without a runbook.
- Separate `/health` for process liveness and load balancers from `/ready` for dependency readiness and orchestration. A health check does not query the database.
- Add a new metric or alert in the same change as the feature it observes. Do not defer monitoring to an unspecified later task.