Re: Monit ignoring Memory amount limit

[William Cotton]

You know what, monit IS taking care of it! What I should do now is change the amount of time in between checks, from 180 seconds to something like 60 or 30...

I guess I'm surprised to find out that in the 15 minutes between 5 cycles, that a single mongrel process can balloon so quickly...

*slaps forehead*

Thanks everyone!

William

On 10/2/07, Martin Pala <address@hidden > wrote:

Try to enable monit verbose mode (using -v option) and check logs ...
it logs/reports the memory usage for each cycle. How does your monit
configuration describe the mongrel processes? Is the process with the
700MB usage related to the monitored parent? (the one with pidfile
matching the monit configuration).

Martin

On Oct 2, 2007, at 8:15 PM, William Cotton wrote:

> Yeah, they are definitely staying above the limit, growing to some
> obscene sizes as well... I've seen one mongrel process get over a
> 700MB!
>
> On 10/2/07, Jan-Henrik Haukeland <address@hidden> wrote: On 2.
> okt. 2007, at 19.09, William Cotton wrote:
>
> > I am monitoring a bunch of mongrel processes, including in their
> > monit configuration:
> >
> > if totalmem is greater than 150.0 MB for 5 cycles then restart
> >
> > which, when I check on it's status, shows:
> >
> > If greater than 153600 5 times within 5 cycle(s) then restart else
> > if passed 1 times within 1 cycle(s) then alert
> >
> > However, I frequently find that processes have swelled to well
> > beyond that limit with no restart. Is there something that I'm
> > missing here?
>
>
> Have you checked that processes stay above 150 MB for the entire
> duration of 5 cycles? If not, the if-test would not trigger.
>
>
>
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general
>
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general



--
To unsubscribe:
http://lists.nongnu.org/mailman/listinfo/monit-general