Log work against a Jira Cloud issue and control how the remaining estimate is adjusted

Documented steps

1. Confirm time tracking is enabled on the site — the worklog endpoints error out otherwise.
2. Add a worklog: POST /rest/api/3/issue/{issueIdOrKey}/worklog with Content-Type: application/json.
3. Body: {"timeSpent":"3h 20m","started":"2026-08-01T09:00:00.000+0000","comment":{"type":"doc","version":1,"content":[{"type":"paragraph","content":[{"type":"text","text":"Investigated the regression"}]}]}}
4. Supply EITHER timeSpent (a string like '3h 20m') OR timeSpentSeconds (an integer), never both — supplying both is rejected.
5. Format 'started' exactly as yyyy-MM-ddTHH:mm:ss.SSS+0000 — milliseconds are required and the offset must be numeric (+0000), not 'Z'. This is the most common 400 on this endpoint.
6. Control the remaining estimate with the adjustEstimate query param: 'auto' (default, reduces remaining automatically), 'leave' (do not touch it), 'new' (requires newEstimate, e.g. ?adjustEstimate=new&newEstimate=2d), 'manual' (requires reduceBy, e.g. ?adjustEstimate=manual&reduceBy=1h).
7. Add &notifyUsers=false when back-filling worklogs in bulk to avoid spamming watchers.
8. Read worklogs back: GET /rest/api/3/issue/{issueIdOrKey}/worklog?startAt=0 — maxResults defaults to 5000 here. Narrow with startedAfter and startedBefore, which take UNIX millisecond timestamps.
9. Success on create is 201 with the created worklog object.