Mean Time to Recovery (MTTR)

Mean time to recovery (MTTR) is the average duration it takes to restore a service after a failure occurs. It is measured from the moment an incident begins, or is detected, to the moment normal service is restored.

How It Works

To compute MTTR, a team records the duration of each incident over a period and averages them. The metric can be broken down to expose where time is lost: time to detect, time to acknowledge, time to diagnose, and time to repair. Reducing MTTR usually means improving detection through better observability, speeding response through clear runbooks and on-call practices, and enabling fast mitigation such as automated rollback. Because a few long incidents can skew the average, teams often examine the distribution rather than the mean alone.

Why It Matters

Failures are inevitable, so how quickly a team recovers is often more important than how often it fails. MTTR captures that resilience in a single number, making it easier to track whether operational practices are improving. It is one of the four key DORA metrics used to assess software delivery performance, alongside deployment frequency, lead time for changes, and change failure rate. A low MTTR signals that a team can absorb failures with minimal user impact, which in turn supports a faster pace of change.

Related Terms

MTTR measures the effectiveness of incident management and on-call response, is reduced by lessons captured in postmortems, and is a DORA metric alongside deployment frequency.