[plug] nagios / services depending on local host

Adam Hewitt ahewitt at theozhewitts.com
Wed Jul 30 15:23:30 WST 2008


Hi All,

Sorry if this email comes through twice, I sent it this morning and it
never showed up

I have spent hours reading the Nagios documentation and I am quite
confused as to how to go about fixing this problem:

I have a bunch of servers being monitored successfully, however if a
server goes down Nagios will continue to run checks against the
services on that server. Each of these checks then have to timeout and
the scheduling queue ends up going through the roof. Is there an easy
way for me to prevent this from happening, ie. if the host is
registered as being down then only check the host and cancel all
service checks on that host?

I think I can do this by setting up each service as dependent on the
host, but that's a *lot* of extra config for something that seems like
it should work out of the box.

oh, its Nagios 2.9 btw.

While I am here, I may as well ask if there are any performance
tweaking that people can recommend for nagios as well. One thing I
have noticed which is a bit weird is that services checks don't seem
to happen in parallel 99% of the time. I look at the scheduling queue
and it shows 10-20 checks scheduled to be run in the same second, but
a `watch -n 1 "ps uax | grep ^nagios"` only shows one being executed
at a time. The max_concurrent_checks value is set to 0 (unlimited)
which I have confirmed by running nagios -s....so something seems to
be preventing them from running concurrently.

Thanks.

Adam.



More information about the plug mailing list