Jobs and schedules

Admin, Fleet, Jobs is the central view of background work. It includes instance-wide jobs, work that runs once per host, and older unmanaged jobs that are visible but not yet controllable through the common scheduler.

What the table shows

For each job you can see its scope, whether it is enabled, its schedule, last and next run, current state, and latest result. Expand a host-scoped job to see the state of each host independently.

Schedules may be periodic, limited to a time window and selected days, event driven, or manual only.

Run now

Use the actions menu and choose Run now. A disabled job cannot run, and a job that is already active refuses a second overlapping run rather than queuing one silently. Host-scoped jobs ask which host should receive the run.

Edit a schedule

Managed schedules can be enabled or disabled and their interval or execution window changed from the page. Some schedules are locked by an environment variable; the UI names that variable and disables the fields it controls. Remove the environment override and restart before expecting an admin change to win.

Unmanaged jobs are listed for visibility but have no editable common schedule.

Run history

Open a job to see bounded, newest-first run history. The result distinguishes a successful run, a failure, a skipped run, and an active run. Use the recorded message and timestamps before searching raw container logs.

Run now can change external state

Jobs do real work: they can scan libraries, install or inspect images, remove old telemetry, or perform host maintenance. Read the job name and target before starting it manually.