[icinga-devel] Make sockets non-blocking

Hiren Patel hir3npatel at gmail.com
Thu May 6 14:24:35 CEST 2010


On Thu, 6 May 2010 10:59:46 +0200
Christoph Maser <cmr at financial.com> wrote:

> > >
> >
> > one suggestion I had, was to start sending alerts to the admin address configured in icinga.cfg.
> > since this would classify as a critical issue, and logging is no longer possible..
> > just my 2c.
> 
> That should be done for sure, but should monitoring continue and throw
> away results or not?
> 

my opinion, it should not.
I think it should exit, causing at least the frontend to fail and get some attention from someone monitoring the frontend.
that being said, I also think default config should have load/disk checks for localhost, set to mail the admin address for when disk/load gets critical.
this should cover for a prewarning on low resources, and then cause action for when resources are out. I don't think icinga should continue to function when resources are out, because to me, it could make it look like everything is normal/ok to an admin or monitoring team, yet behind the scenes logs/historical data is being lost. IMHO, rather fail right away and have the problem sorted out.

-- 
Hiren Patel <hir3npatel at gmail.com>




More information about the icinga-devel mailing list