Skip to content

    MTTR meaning: what MTTR is, how to calculate it and how to reduce it

    Reference baseline: 7 October 2026. Confirm current availability and terms with each vendor.

    MTTR is the average time it takes to restore a system after a failure. The R varies: recovery, repair, resolve or respond.

    Teams that compare MTTR without agreeing which one they mean are comparing different clocks. Before you set a target, write down where your clock starts and where it stops.

    Watch a recorded Operate investigation

    The four meanings of MTTR

    Where each MTTR clock starts and stops

    MTTR
    Mean time to recovery (or restore)
    Clock starts
    The failure begins
    Clock stops
    The service works again for users
    MTTR
    Mean time to repair
    Clock starts
    Work on the fix begins
    Clock stops
    The fix is in place
    MTTR
    Mean time to resolve
    Clock starts
    Your agreed incident start
    Clock stops
    The incident meets your agreed resolution criteria
    MTTR
    Mean time to respond
    Clock starts
    The alert fires
    Clock stops
    Someone starts working on it

    How to calculate MTTR

    MTTR is the total restore time across incidents, divided by the number of incidents.

    An illustrative example with round numbers: four incidents last month took 30, 45, 60 and 225 minutes to restore. The total is 360 minutes. Divided by 4 incidents, MTTR is 90 minutes.

    Why the average MTTR misleads

    In the example above, three of the four incidents were restored in an hour or less, but the one long incident pulls the mean up to 90 minutes. Outage durations have long tails, so a mean is dominated by a few bad incidents. The median of the same four incidents is 52.5 minutes.

    Track the median and a high percentile alongside the mean to see how far apart they get.

    MTTR vs MTTA vs MTTD vs MTBF

    MTTR and the related reliability metrics

    Metric
    MTTD
    Stands for
    Mean time to detect
    What it measures
    Failure start to detection
    Metric
    MTTA
    Stands for
    Mean time to acknowledge
    What it measures
    Alert to a person acknowledging it
    Metric
    MTTR
    Stands for
    Mean time to recovery (or repair, resolve, respond)
    What it measures
    Time to restore, by your chosen definition
    Metric
    MTBF
    Stands for
    Mean time between failures
    What it measures
    Average working time between one failure and the next

    MTTR in DORA metrics

    DORA's software delivery metrics no longer use the name MTTR. The closest one is failed deployment recovery time, which DORA defines as "the time it takes to recover from a deployment that fails and requires immediate intervention" (dora.dev, checked 7 October 2026). It covers deployment failures only, not every incident.

    What is a good MTTR?

    There isn't one number. A good MTTR depends on which definition you use, how severe the incidents are and what your users tolerate. Set the target from your own median and from your service-level objectives, not from another company's figure. Our guide on what a good MTTR is and how to set your own target walks through it.

    How to reduce MTTR: fix diagnosis first

    Our view is that diagnosis, finding out what broke, is the part of the restore clock most worth shortening: faster access to logs, database state and recent changes, and a root cause with evidence instead of a guess. Write the steps of a few recent incidents down with timestamps and you can see where your own clock goes.

    That is the part an AI SRE automates. Operate reads connected logs, records and code, cross-checks proposed causes and can draft a patch for engineer review. Read more in how to reduce MTTR.

    Frequently Asked Questions

    Is MTTR a KPI?

    Yes. Many teams use MTTR as a key performance indicator for incident response and reliability. It works best with an agreed definition of which MTTR you track, and alongside the median and a high percentile, because a few long incidents can move the average a lot.

    What is a good MTTR value?

    There isn't a universal good value. It depends on which MTTR you measure, how severe the incidents are and what your users tolerate. Set the target from your own median restore time and your service-level objectives, then work to bring it down, starting with the time spent on diagnosis.

    What is the difference between MTBF and MTTR?

    MTBF, mean time between failures, measures how long a system runs between one failure and the next. MTTR measures how long it takes to restore the system after a failure. MTBF tells you how often things break, and MTTR tells you how quickly you recover.

    Is MTTR the same as rto?

    No. RTO, the recovery time objective, is a target: the longest downtime you are willing to accept, set in a continuity or disaster-recovery plan. MTTR is a measurement of how long restoring actually took, on average. Compare individual recovery durations with the relevant RTO; an average below the target can still hide incidents that exceeded it.

    What is MTTR in maintenance terms?

    In maintenance, MTTR usually means mean time to repair: the average time to repair a failed piece of equipment and return it to service, from the start of repair work to the end. Software teams more often mean mean time to recovery or to resolve, so say which one you use.

    Sources

    Checked on 7 October 2026.