Nagios Performance Issues

[Mika Fischer]

* Babak Pasdar <bpasdar at pasdar.com> [2004-02-18 01:23]:
> We have recently implemented Nagios 1.2.  We have a total of 620 services on 200
> hosts being monitored.  Services include network port monitoring, snmp checks via
> snmpget and server monitoring via NRPE.
> 
> Our challenge is that Nagios is slow (Extremenly Slow) on sheduled checks.  For
> example at 11:00pm the nagios scheduling queue is still waiting to process scheduled
> service checks from 7:00pm.  Also it takes so long for services or hosts which are
> up to be recognized, probably as a by product of the above.
> 
> Now the system, a dual 2.4 Ghz. P4 Xeon with hyperthreading and 1 Gig Ram always
> seems to have plenty of idle CPU, but sometimes the load level goes as high as an
> insane 1300 with extremely high process counts (7000 or more), when we set the
> checks to be more aggressive.  What I dont understand is that at a load level of
> 1300 the system shows 50% or more idle CPU.

We were having the exact same problem.

> How can we get more aggressive checking and eliminate the backlog of checks without
> driving the sytem loads and process count so high.  In particular what settings
> should we play with for more consistent performance.
> 
> My guess is that we have to play with the following:
> 
> inter_check_delay_method=s
> service_interleave_factor=s
> max_concurrent_checks=0
> service_reaper_frequency=30
> sleep_time=1
> 
> Am I right?

I would start with setting service_reaper_frequency to 1 and
max_concurrent_checks to around 100 or so (also check what nagios -s suggests.

Also note that the snmp-Plugin in its default configuration is a real resource
hog. Use the option "-m :" to prevent snmpget from parsing all the MIB files
and use numeric OIDs.

> If so what formula does Nagios use to do it's smart checking?

Please check th docs for that. But we didn't even need to mess with those

The configuration above works quite well for us checking about 450 Services on
85 Hosts on a P4-2.4GHz (Hyperthreading disabled) with 512 MB of RAM.

HTH,
 Mika
-- 
-------------------------------------------------------------------------
Mika Fischer                                          mika.fischer at to.com
Thinking Objects Software GmbH                  phone:  +49.711.88770.154
Lilienthalstraße 2                              mobile:   +49.175.5249209
70825 Stuttgart-Korntal, Germany                fax:    +49.711.88770.449
-------------------------------------------------------------------------


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id56&alloc_id438&op=click
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null