MTTR in Jira: recovery time
Jira Cloud, failures and fix versions.
MTTR, mean time to restore, is how long it takes to recover when a change causes a failure. DORA now calls it failed deployment recovery time, and measures it as a median. This page says how to measure it from work items in Jira, then how Changeline shows it.
What Jira does by itself
Jira has no recovery time report. What it keeps is each failure as a work item
with a created date, a resolved date and, for a bug, the Fix versions that says
which release fixed it. By hand, a JQL search such as
issuetype in (Bug, Incident) AND resolved >= -90d lists the failures resolved
in the period. Export them with their created and resolved dates, subtract, and
take the median or the mean in a spreadsheet. Incidents often live in a service
space apart from the code’s, such as a Jira Service Management space, so the
search names those spaces.
A failure not yet resolved has no resolved date, so a search by resolved date leaves it out until it is.
What Changeline does
Changeline is our app. For each failure recovered in the period it takes the creation to the release that ships its fix, named in its Fix versions, else to its Done, and shows the median, such as 6h 20m, with the count under it and a lozenge for better or worse than the days before. A median moves less than a mean when one failure ran long. A failure is a work item of a failure type that points at a release: Bug and every work type whose name contains Incident by default, and a Jira admin can name others and add spaces to also read failures from. An incident is no longer a measure of its own: it is a failure type.
The Failures tab lists each failure with its space, when it was created, the release that ships its fix and how long it took. Failures not yet recovered are listed apart and named in an amber box, so a long one is not forgotten while it is open. Changeline only reads Jira and changes nothing in it.
Recovery time here is a work item’s creation to its fix, not an outage’s start to its end, so it is as true as the failure work items and their Fix versions are kept.
See Failures and recovery time and DORA metrics in Jira. Back to Changeline.
DORA here is the research programme at Google behind failed deployment recovery time and four other measures, not the EU’s Digital Operational Resilience Act. DORA is a trademark of Google LLC, and Changeline is not made or endorsed by Google.
