Stressing Nagios

[abc]

Hi everybody.

I am looking to test Nagios to see how it copes with the stress of monitoring numerous services. The goal is to get an idea if I will need to use a distributed setup. I have a test server that is currently monitoring 15 or so services over 3 hosts. I want to increase the number of services being monitored by incrementally increasing the number of services Nagios monitors. I was thinking of appending the same services I am currently monitoring to the end of the nagios.cfg file and then starting nagios again, appending the same services to the end of the nagios.cfg file... until the nagios tools begins to show signs of stress. Hopefully this will give me a good idea of what Nagios is able to cope with. Does anybody have any idea of how to do this?

[Box293]

What you are doing is logical. If you are appending the same services, you'll need to make sure they have a different service description so they are treated as separate services and not ignored because they are duplicates.

Hopefully this will give me a good idea of what Nagios is able to cope with

How are you objectively measuring this?

Also, I've done testing like this in the past. What you want to do is take a host or two offline so all the services enter a warning and critical state and trigger event handlers and notifications. A lot more resources get used when services start retrying checks at the retry_interval instead of the normal check_interval.

Also, it depends on the plugin being used. For example nrpe checks that are executed remotely consume less resources than snmp or perl checks that run on the Nagios host.