MTTR, or mean time to repair, is the average time it takes your team to restore service after an incident. The MTTR calculator turns a list of incident durations into that average, and shows which part of the response — detection, response, or repair — is costing you the most time.

A single bad outage tells you little about your team. An average over a month or a quarter tells you whether recovery is getting faster or slower, and whether a change you made actually helped. That is what MTTR is for.

How MTTR is calculated

MTTR is the total repair time for all incidents divided by the number of incidents. If five incidents took 30, 45, 60, 90, and 120 minutes to repair, the total is 345 minutes and the MTTR is 69 minutes.

Repair time is usually split into three phases. Detection is the time from when the incident starts until someone notices. Response is the time from detection until work begins. Resolution is the time from work beginning until the service is restored. MTTR is the sum of all three, so a slow alert can hurt as much as a slow fix.

This calculator works from the durations you enter. Add each incident’s total repair time, or enter the three phases separately to see where the time goes. The average is what matters for reporting; the split is what matters for improving.

Worked example

A team handles four incidents in a month. The repair times are 20 minutes, 35 minutes, 50 minutes, and 3 hours 15 minutes (195 minutes). The total is 300 minutes, and with four incidents the MTTR is 75 minutes.

The average hides an important detail: three of the four incidents were fixed in under an hour, and one took more than three. That long incident may have been a different kind of problem — a database failure rather than a service restart — and it may deserve its own review rather than being averaged away.

If the team splits the 195-minute incident into phases, the picture sharpens further. Detection took 40 minutes because the alert went to an inbox nobody watched overnight. Response took 25 minutes. Resolution took 130 minutes. The biggest win is not a faster fix; it is a better alert.

How to use the result

Report MTTR by severity, not as one number for everything. A critical outage and a minor ticket should never share an average, because the minor tickets will hide the critical ones. Most teams track P1 and P2 separately, and that is enough.

Compare MTTR across months to see trends, but look at the median too. The mean is pulled up by rare long incidents, while the median shows what a typical incident looks like. If the mean is falling but the median is not, your typical incident is unchanged — you just had fewer disasters.

Use the phase split to choose your next improvement. Slow detection means better monitoring and alerting. Slow response means clearer on-call rotation and escalation. Slow resolution means runbooks, practice, and access to the right systems. Each fix is different, and the split tells you which one to fund first.

Note: Do not set an MTTR target your team can only hit by closing incidents early. If the target creates pressure to mark things resolved before they are fixed, the number improves and the service gets worse. Pair MTTR with a reopened-incident count.

MTTR and the cost of downtime

Every minute of MTTR is a minute of downtime, and downtime has a price. Multiply your MTTR by the hourly cost of an outage to see what slow recovery is worth each month. The downtime cost calculator does that math, and it often turns a vague we should be faster into a budget line with a number attached.

Frequently asked questions

What is MTTR?

MTTR is mean time to repair: the average time from when an incident starts until the service is restored. It is the most common measure of how fast an IT team recovers from outages.

When does the MTTR clock start?

It starts when the incident begins, or when it is detected if the start time is unknown. Most teams measure from detection, because that is the first timestamp they can prove.

Does MTTR include the time waiting for a vendor?

Yes, if you measure end-to-end repair time. Some teams track a second number that excludes waiting on third parties, so they can separate their own speed from a vendor's.

What is a good MTTR?

It depends on severity. Many teams target under 1 hour for critical incidents and under 4 hours for high severity. The right target is one your team can hit most of the time, not one that looks good on a slide.

How is MTTR different from MTBF?

MTTR measures how fast you recover after a failure. MTBF measures how long a system runs between failures. Together they describe both halves of reliability: how often things break and how fast you fix them.