monit needs to monitor stopped processes and kill them if they come back alive.
Issue #564
new
We have a cluster and use monit to "group" several processes together and start/stop them when needed.
Recently, we noticed that not all processes in a group are stopped. Monit sends out a stop signal, but for some reason not all processes die. Monit is not doing a second attempt at stopping them.
To be truly HA compliant, monit needs to kill processes which are marked as stopped even if they accidently start again by user or don't get killed in the first request.