Job lifecycle

run creates a persisted job and normally observes it briefly. The returned job_id identifies the same job for status, wait, tail, kill, and restart.

1. Run a command and retain the job_id.
2. Use status for current state.
3. Use wait for a bounded observation window and terminal exit code when available.
4. Use tail for bounded, repeatable log excerpts.
5. Use kill only to explicitly cancel a running job.

Initial observation ending does not stop the managed command.