List running job claims
GET /admin/jobs/workers
Returns running job claims and their lease details.
You need to be signed in with system.jobs.read or system.admin for the whole install; a permission held only for a team doesn’t count.
Each item is a job whose stored status is running and that has a claimed_by value. It isn’t a list of worker processes: idle workers are absent, one worker ID can appear more than once, and a claim can outlive its worker. Claims from in-process and external workers appear together.
Rows are sorted by claimed_by, then claimed_at, with no further tie order. The response isn’t paginated or limited to the local pool’s size. With no matching claims, items is empty.
lease_stale is true when the recorded expiry is earlier than the database time used by the query. The watchdog can requeue a stale lease, but this read doesn’t reclaim it or guarantee recovery. The standard pool checks for expired leases every 30 seconds; worker heartbeats, database errors and shutdown can affect what happens next. Call the endpoint again for a fresh database view.
Authorizations
Section titled “Authorizations ”Responses
Section titled “ Responses ”Running job claims with worker identifiers and lease information.
Running job claims visible in the database.
object
One item per running job with a recorded worker ID. Idle workers are absent and IDs can repeat. Empty when no claims match.
A running job claim, including its recorded worker identifier and lease.
object
Worker ID stored with the claim. The same ID can appear on several returned jobs.
UUID of the claimed job.
Type of the claimed job.
Stored queue priority; lower numbers are considered first when claiming eligible work.
Number of claims recorded for this job since creation or administrative requeue.
When the claim was recorded. Left out when absent.
Recorded expiry used by lease recovery. Left out when absent.
True when the recorded lease expiry is earlier than the database query time. The claim is then eligible for watchdog recovery, but that doesn’t mean recovery has happened or that the worker is dead.
Authentication required, missing, or invalid
object
Human-readable error summary
Example
the request could not be completedExample
{ "error": "authentication required: sign in and retry with a valid session or API token"}Authenticated but missing required capabilities
object
Human-readable error summary
Example
the request could not be completed