Mean time to restore
Stability half of the DORA pair, alongside CFR. Detection to restoration.
OW
Oliver Wakefield-Smith
Founder, Digital Signet. Built siemcost / monitoringcost / itbudgetcalculator.
Updated 21 Jun 2026
Direct answer
Mean Time to Restore (MTTR) is the time from incident detection to service restorationGoogle Cloud MTTR Guide2024. 2024 DORA elite teams sit under one hour; medium teams in the 1-24 hour band DORA 2024 Report2024. MTTR is not MTBF or MTTD; the terms are commonly conflated.
MTTR vs MTTD vs MTBF
MTTD = detection time (when the alert fires). MTTR = restoration time (when service is back). MTBF = mean time between failures. DORA cares about MTTR; SRE practice cares about all three plus MTTM (mitigation, often used as a proxy for restoration).
Sourcing MTTR from PagerDuty
PagerDuty Incidents API exposes created_at and resolved_at per incidentPagerDuty Incidents API2026. MTTR = median(resolved_at - created_at) over the window. See /signal-sources/pagerduty for the full pipeline.
Continue reading