Calculate mean time to recovery from incident detection to full resolution.
MTTR (Mean Time To Recovery/Repair) is the average time it takes to restore service after an incident begins: MTTR = total_downtime ÷ number_of_incidents. Projecting the observed incident rate forward to a full year gives a sense of the total annualized downtime impact if current reliability trends continue unchanged — useful for prioritizing incident response investments against their expected payoff.
MTTR = totalDowntime / numberOfIncidents
No — MTTA (Mean Time To Acknowledge) measures detection-to-acknowledgment, MTTR measures acknowledgment (or detection) to full resolution, and MTBF (Mean Time Between Failures) measures the gap between incidents, not their resolution speed.
It varies heavily by system criticality — high-availability services often target MTTR under 30-60 minutes, while less critical internal tools may tolerate several hours.
Faster detection (better alerting), clearer runbooks, automated rollback/remediation, and practiced incident response (game days, postmortems) are the highest-leverage levers.
Yes — MTTR is an average, so a single multi-hour outage can dominate the metric. Many teams also track median recovery time or a 90th-percentile MTTR alongside the mean to catch this.