|From:||Smith, Graham - Computing Technical Officer|
|Date:||Mon, 20 Feb 2017 16:06:54 +0000|
I’m successfully using monit on a number of servers for a variety of diverse tasks.
But now I’d like to monitor some desktops and other devices that may be down or unplugged
by a user – but not indefinitely.
A desktop may be not used for up to about 4 days and that would be normal enough if
classrooms aren’t full for some groups and 2 of those days are the weekend.
If it is not used beyond about 4 days, that it raises suspicion that there may be a
failure worthy of investigation. In short I need a more “fuzzy” uptime check.
I can use the cron section so I don’t get alerts outside of term time and that’s useful.
But the maximum limit of 64 cycles isn’t really enough. With a 30 second cycle I’d need something like:
check host Lab220-01 with address 10.127.220.1
if failed ping with timeout 10 seconds for 11520 cycles then alert
I could script something externally to keep a count of length of outage producing a score similar to how e-mails are scored for probability of
spam content. Then poll that in monit based on exit status. But that seems an inelegant solution. Anybody using monit in this way?
|[Prev in Thread]||Current Thread||[Next in Thread]|