Skip to content

.tip

· one of the small things

Every alert carries the link to its runbook

At 3am the responder needs to know what the alert means, what to check first and what to do, in that order, without having to search.

annotations:
  summary: "checkout p99 above 1.5s for 5m"
  runbook: https://wiki.example.com/runbooks/checkout-latency

If an alert has no runbook, write one or delete the alert. An alert nobody knows how to act on is just noise with a pager attached.

observabilitysre