Re: hierarchical network monitoring

[Martin Pala]

Hi,

you can unmonitor the services which are behind the failed switch.

For example if switch is monitored via availability of switch's network 
management interface on port 1234, then put all depending services into some 
group and disable monitoring of this group on switch0 error, else if recovered 
then enable monitoring of it.

pseudo configration:
--8<--
check switch0 with address ...
   if failed port 1234 then exec "/bin/monit -g switch0 unmonitor" else if 
succeeded then exec "/bin/monit -g group_switch0 monitor"

check node0
  group switch0

check node1
  group switch0

check node2
  group switch0
--8<--


Best regards,
Martin



On Sep 10, 2010, at 12:16 PM, Kamenik, Aleksander wrote:

> Hi,
> 
> I was wondering if one could setup monit to ignore failed hosts (and/or 
> services) that are behind a failed router.
> 
> For example: http://pastebin.com/S2vF0g5Z
> 
> If switch0 goes dead, I don't want the alerts for node0-2. If fw0 fails, I 
> don't want the alerts for all the stuff behind it. Though it would be nice it 
> would mention in the alert for the failed network device that there are N 
> devices behind it that are probably unreachable now. Assuming no redundant 
> route.
> 
> Point being, I don't want 10 SMS messages when one is enough.
> 
> Is it possible create something like this using groups?
> 
> Regards,
> 
> Aleksander Kamenik
> System Administrator
> Krediidiinfo AS
> an Experian Company
> Phone: +372 665 9649
> Email: address@hidden
> 
> 
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general