Hanging requests #710
Description
Sometimes a request is not terminated correctly and stays in the system for days. This reduces resources we have, requires manual clienup and creates huge problems on platforms like NRP. There if a pod stays for a long time and not uses cpu, it is counted as a violation and that user can't create new jobs.