MTTR (mean time to recovery)
Mean time to recovery (MTTR) is the average time between a failure starting to affect users and the service working normally again. It measures how fast a team detects, diagnoses and fixes problems. It was one of the original four DORA metrics for software delivery performance.
How it works
Add up the recovery time of every incident in a period and divide by the number of incidents. Four incidents that lasted 20, 45, 10 and 85 minutes give 160 / 4 = 40 minutes MTTR.
The "R" is used loosely. It can mean recover, restore, repair or resolve, and each starts and stops the clock at different points. Write down the definition your team uses. DORA has since moved from MTTR to failed deployment recovery time, which counts only failures caused by deployments.
Why it matters in production
Failures will happen, so recovery speed often matters more than failure count. MTTR improves with better alerts, clear runbooks, fast rollbacks and practised on-call engineers. Because it is an average, one long outage can skew it, so look at the individual incidents too.
See also
On the One2N blog
Further reading
