I have a group of machines on the same subnet running monit, where
several of the machines are "check host" each other. These machines
don't generate FPs amonng themselves.
I have another machine on a remote site running monit that "check
host" the group of machines, with
check host hostname-ALIVE with address hostname
if failed icmp type echo count 5
with timeout 1 seconds 3 times within 3 cycles
then alert
The average ping time from remote monit to the group is 65 ms.
The remote monit sporadically generates false alerts for only one
machine at a time in the group (not alerts for the entire group),
with the FP moving among the group. When the remote monit sends an
alert for machine x, the other machines in the group also monitoring
machine x send no alert.
I'm talking 2 or 3 FPs/day. Not a disaster but does anybody have
any suggestions how to kill these FPs? I could up to "count 50 with
10 times within 10 cycles", but I'd rather understand why the above
doesn't suffice.
thanks
Len
--
To unsubscribe:
http://lists.nongnu.org/mailman/listinfo/monit-general