Retention & the watchdog
Keep storage bounded and jobs honest.
Retention
A daily sweep prunes artifacts older than each policy’s retention window, and old finished job rows are pruned too (the latest run per policy is always kept) so the history stays manageable.
The watchdog
Running jobs heartbeat while they work. A periodic, multi-instance-safe reaper fails any job whose worker stopped responding or that exceeds a max runtime — so a crash never leaves a job stuck "running" forever.
Try it yourself
Open the console and put this into practice.