General Nagios question regarding resources/CPU usage, etc

[xpac]

I've gotten Nagios rolled out into our production environment, and so far I'm very happy with it. That said, yesterday I was running top on the Nagios server and saw that at times nagios would pop up to the top using 99-100% CPU briefly, and then it would fall to the bottom of the list. This doesn't particularly worry me, as there are 24 cores in each box so that only represents basically one core at those moments, just wondering if maybe it was performing a check at those moments where it jumped up there?

Also, what, if any, performance hits should I be concerned with having the Nagios server running on one of the production boxes. Its a very small environment (for now anyway), less than 10 boxes are currently being monitored if that helps. Just trying to think ahead and make sure we don't run into any issues moving forward from a performance standpoint.

[eloyd]

Nagios load usage depends on specific service check frequency, what the service checks do, what sort of notifcations you have, state stalking, logging, and a host of other factors. 24 cores is...um, probably a LOT for Nagios. We run our internal Nagios box on a VM with 512 MB of RAM and one core, for instance. Wait. I may be lying. We might have just bumped it up to 1GB of RAM and 2 cores for doing demos at the Nagios Conference last week.

I am guessing that you will have zero problems with Nagios interfering with your production boxes. One of our previous clients (a very large e-commerce site) had Nagios installed on the same box that was the push-off point for their promote-to-production path (too many "p"'s in that sentence). No problems there.

Shameless sales pitch - my company specializes in providing Nagios consulting services. I'd be glad to hook you up with a phone call if you want to discuss further.

[abrist]

Hammers and nails eloyd.

[eloyd]

I'm just jealous that he has 24 cores.

[abrist]

And the earth is fine with just one! I think OP is greedy . . .
Seriously though, did eloyd's suggestions work for you xpac?

[xpac]

Sorry guys I forgot to check back.

Yes that is fine, actually I have 6 boxes each with 12 physical cores but with hyperthreading enabled each has 12 virtual cores as well. I just chose one of the boxes of our database cluster to be the Nagios server, since well I don't have anything else to put it on at the moment and wanted to make sure I didn't kill our production environment

That said, I REALLY want at least one dedicated management server, that is physically separate from our production cluster, but it looks like that will be put on the back burner for now because based on this thread I shouldn't need to worry for now

Unfortunately the guy who hired me, and set up the environment initially quit so I'm gonna have to figure out how to convince the CEO to invest in management infrastructure which right now seems more like a luxury than necessity.

[tmcdonald]

I'm gonna have to figure out how to convince the CEO to invest in management infrastructure

You are a far braver soul than I.

Are we clear to lock the thread?

[eloyd]

Unfortunately the guy who hired me, and set up the environment initially quit so I'm gonna have to figure out how to convince the CEO to invest in management infrastructure which right now seems more like a luxury than necessity.

My company offers Nagios consulting and management services if you want to outsource that problem to us.

[tmcdonald]

Gonna lock this since it looks like you know what to do.