I've "accidentally" created a health check that failed to terminate. It basically failed into an infinite loop. This prevented the system from recovering via rollback:
[ ***] Job greenboot-healthcheck.service/start running (59min 59s / no limit)
[ **] Job greenboot-healthcheck.service/start running (1h / no limit)
[ *] Job greenboot-healthcheck.service/start running (1h / no limit)
[ **] Job greenboot-healthcheck.service/start running (1h 1s / no limit)
etc.
Would it be possible to add a greenboot configuration option to place a maximum limit to how long the checks can run? If the limit is hit then greenboot could mark the check as failed an rollback.