My idea is to allow cancelling an stuck run with all jobs done, in case such a bug happens again to not require admin commands. Related #35832 Co-authored-by: Zettat123 <zettat123@gmail.com> Co-authored-by: silverwind <me@silverwind.io>