I intend to write a process watchdog that runs out of cron or similar that watches all of the core processes denoting things like run state, VMstat, Inet operations and reports it to a file or external host.

If a process dies that is deemed core it restarts it. Is an already written watchdog existent that can address my needs?

  • See: upstart, systemd, and supervisord.
    – jordanm
    May 28 '13 at 13:11
  • I think Zabbix could help you. It's used for monitoring and can trigger alerts and takes actions if you configure it to do so. It's quite a big thing though...
    – lgeorget
    May 28 '13 at 13:12
  • @Rohan systemd and supervisord have decent re-spawning capabilities (with logs) but don't report resource utilization by themselves. You could add it to a cgroup and then just write a cronjob that emails the stats for the cgroup wherever you're wanting it to go, though.
    – Bratchley
    May 28 '13 at 14:18
  • @Rohan & @jordanm, I have tried with supervisord, not able to controll properly. Here is my question-->unix.stackexchange.com/questions/358522/… Apr 13 '17 at 7:33