Re: Monitoring a network interface

[Martin Pala]

Hi,

we discussed such feature in the past - the syntax of these triggers 
could be something like this:

 if <x> <event> within <y> cycles then <action>

where:
.. x      = number of given event type occurences
.. event  = event type trigger restriction/specification
.. y      = number of cycles
.. action = what to do


The event engine could be extended to provide this functionality - 
famous quotation: 'it needs only to be implemented' ;)

Martin

Chris Boot wrote:
> Hi all,
>
> I've recently begun to use monit to monitor my server, and boy does it 
> come in handy! Cheers for this great little piece of software.
>
> Now I'd like to set up a monitor to test network connectivity. Ideally, 
> what I'd like to do, is to ping and HTTP connect to a couple of hosts 
> and if they're down, bring down and bring back up again my external 
> interface, eth1. If they are brought down say 5 times in 5 cycles and 
> it's still not working, reboot. At the moment I have:
>
> check host debian with address www.debian.org
>        group connectivity
>        start program = "/sbin/ifup eth1"
>        stop  program = "/sbin/ifdown eth1"
>        if failed icmp type echo with timeout 15 seconds then alert
>        if failed port 80 proto http then restart
>
> check host apple with address www.apple.com
>        group connectivity
>        start program = "/sbin/ifup eth1"
>        stop  program = "/sbin/ifdown eth1"
>        if failed port 80 proto http then restart
>
> Now this just restarts the interfaces as soon as a problem is detected, 
> which works but is a bit clunky, as you might imagine.  Is there any way 
> to make, for example, something like the following (that monit doesn't 
> like)?
>
> check host debian with address www.debian.org
>        group connectivity
>        start program = "/sbin/ifup eth1"
>        stop  program = "/sbin/ifdown eth1"
>        if failed icmp type echo with timeout 15 seconds then alert
>        if failed port 80 proto http then alert
>        if 3 alerts within 3 cycles then restart
>        if 3 restarts within 20 cycles then exec "/sbin/reboot"
>
> Thanks for any help!
> Chris
>
>
>
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general