- The four commands
- PAUSE accumulates, it does not discard
- RESTART always archives
- Counting the cycles
- Choosing between STOP and RESTART
- Related
A stopwatch is driven by four commands. Which command a Jira event sends is up to the configuration; what each command does to the value is fixed, and is what this page is about.
The state machine itself - the states, and which command wins when two triggers fire at once - is on Timers and Stopwatches - General information.
The four commands
| Command | Effect on the value |
|---|---|
| START | Starts counting. On a paused stopwatch it resumes the current cycle - it does not start a new one and it does not reset anything. |
| PAUSE | Suspends counting. What has been counted so far is accumulated into timeSpent, and the next START continues from there. |
| STOP | Ends counting. The value is final. After a STOP the only command that starts the stopwatch again is RESTART. |
| RESTART | Archives the current cycle and opens a new one from zero. Always. |
PAUSE accumulates, it does not discard
PAUSE is not a reset and not a stop. The elapsed part of the cycle is added to timeSpent, the counter goes quiet, and a later START picks up from that accumulated total.
A cycle can therefore be paused and resumed any number of times: what the stopwatch shows is the sum of the counted stretches, not the wall-clock span from the first start.
This is why a report can offer an average pause time at all: paused time is a quantity the app keeps, not a gap it forgets. See Statistics and averages in reports.
RESTART always archives
RESTART never erases a cycle. It closes the current cycle, files it in the archive, and opens a new cycle starting from zero.
That holds in both ways a restart can happen:
- A live action - an event fires the RESTART command, or an administrator restarts the stopwatch by hand or over the API. The current cycle is stopped, archived, and a new one starts.
- A restore from issue history - when Restore timer and stopwatch by issue history replays an issue’s history and finds a restart in it, it archives the cycle it had built up to that point and opens the next one, the same as the live path does.
Both paths produce the same shape of result: a chain of archived cycles plus one current cycle.
Counting the cycles
The number of restarts an issue has been through is exposed on the value as restartCount.
restartCount = 0- one cycle. Nothing was ever restarted; the current value is the whole story.restartCount = 3- four cycles: three archived plus the current one.
Use it to tell “this took 4 hours” from “this took 4 hours on its fourth attempt” - two very different service stories that look identical if you only read the duration.
Where it is reachable:
- in a value template, to render the attempt count next to the time on the issue;
- as a number of restarts column function in a table report, to group or filter by how often work was reopened.
Choosing between STOP and RESTART
Both end the current measurement; they differ in what happens next.
| You want | Send |
|---|---|
| One final number per issue, no further counting | STOP |
| Each pass measured separately, all passes kept | RESTART |
| Counting suspended and resumed later, as one measurement | PAUSE, then START |
A common mistake is wiring RESTART to an event that fires often - it silently multiplies the cycles on an issue and makes per-issue averages look far lower than the real handling time. If an event should only interrupt counting, send PAUSE.
Related
- Timers and Stopwatches - General information - states and command precedence.
- Stopwatch config - which events send which command.
- Available events (triggers) - the events a command can be bound to.
- Restore timer and stopwatch by issue history - rebuilding the cycles of an existing issue.

