What a fast recovery actually looks like when the runbook, alerts and traces line up.

A five-minute recovery is not luck and it is not heroics. It is the visible result of three things being ready before the incident: a precise alert, a trustworthy trace, and a runbook someone can actually follow.

Minute one: a precise page

The alert did not say "something is wrong." It said which service, which symptom, and linked straight to the relevant dashboard. The on-call engineer knew what they were dealing with before they finished reading.

Minutes two and three: a trustworthy trace

One click from the alert landed on an exemplar trace of a failing request. The failure was obvious in the span tree — a dependency returning errors — so there was no guessing about where the time went.

Minute four: a runbook that works

metrics
| where __name__ == "dependency_error_rate"
| stats rate by peer.service
| sort rate desc

The runbook's first step matched the symptom exactly and gave a copy-pasteable mitigation: shed load from the failing dependency and serve the cached fallback.

Minute five: confirmed recovery

The same dashboard that raised the alarm showed error rate falling back to baseline. Fast MTTR is what you get when every step of the path was built in advance — the recovery is just you walking a path you already paved.