Gauging how much Nagios is needed

[cwscribner]

Hi all.

I was just wondering how to gauge what hardware is needed depending on the number of hosts. Any suggestions on how to calculate hardware to hosts?

[nscott]

There isn't a standard metric for it, however in the http://assets.nagios.com/downloads/nagi ... rmance.pdf doc, it is discussed what the repercussions of checks was on a given machine, which could be used as a yard stick to determine your necessary hardware requirements.

[cwscribner]

I was wondering if I could get a bit more feedback on this topic? I have an instance of XI monitoring ~1600 devices and ~1750 services. Its gotten to a point where it takes forever to check the configs and process any changes I make. Long term, I'll be implementing a distributed monitoring solution but for right now I don't have that option. Please advise when available.

[nscott]

The checking of the configs and making changes is a very disk IO intensive process. There is a server statistics dashlet that comes standard with XI (and defaults to the home dashboard.) is the IO wait time on that absurdly high (higher than 20%)? The readings on the gauges in the Server Statistics dashley can be a good benchmark as to what part of your system is being taxed.

From what you're saying, it might be worth offloading the Nagios XI MySQL database. It would take some IO strain off the hard drive to be sure! If you have an older server sitting around doing nothing it would work as a slave for the Nagios XI with small amount of effort. Also, I know not everybody can afford the time of having a development box, but if you can swing the time, it does make researching better ways for your particular complex situation a lot easier and worry free. Either way, give this doc a look:

http://assets.nagios.com/downloads/nagi ... Server.pdf

[cwscribner]

I have seen the IO wait time showing red several times within the past day or so. The load metrics also spend a small portion of thier time in the yellow. The check time and check latency look pretty low, however. Does that mean the problem is more so with the hardware than with Nagios?

[mguthrie]

I might also suggest running the repair procedure on the mysql databases just to make sure there's not any corruption slowing them down.

http://assets.nagios.com/downloads/nagi ... irdatabase

[cwscribner]

Already done.

[mguthrie]

Here are a few other known tweaks that can be done:
http://assets.nagios.com/downloads/nagi ... p#boosting

[cwscribner]

Hmmm...so it seems that this is in fact a hardware limitation? I've already made as many changes as I can at the moment. I don't have an additional server to offload mysql to right now. Does this mean that I have to add all the devices by hand?

[mguthrie]

Based on the Write Config Tool failing as mentioned in the other thread, I'm concerned there could be other issues on this system. Lets continue that discussion in this thread:
http://support.nagios.com/forum/viewtop ... =16&t=2978