[CentOS] Nagios, getting started

Mon Jun 22 17:58:47 UTC 2015
Chris Beattie <cbeattie at geninfo.com>

On 6/22/2015 6:04 AM, Rob Kampen wrote:> What advantages does the RH/CentOS world have (if any?) over the ubuntu 
> LTS world?

I can't think of any compelling reason to run Nagios on RHEL/CentOS if the rest of your shop is Ubuntu.  If everyone there is familiar with Ubuntu, it'll be easier for them to troubleshoot a problem if they don't have to learn a new package manager at the same time.

> If a vm is okay, what kind of RAM does it need?

A VM is fine.  I have Nagios monitoring 1,800 hosts and 17,000 services on a VM with 3 CPUs, 4GB of memory, and 20GB of storage.

Whatever you do, set up MRTG graphing so you know how well Nagios is performing (http://nagios.sourceforge.net/docs/3_0/mrtggraphs.html) on your monitoring host.

If you have many hosts and services to monitor, look in to the large installation tweaks and other advice in the tuning guide (http://nagios.sourceforge.net/docs/3_0/tuning.html).

Also, Nagios causes constant disk activity temporarily storing and processing check results, so you can use a RAM disk to speed that up.

> I see epel has nagios 3.5.1 with a date of 2013 for CentOS7 along with 
> plugins, is this the version folk use? - as the latest from nagios is 4.0.8

I still use 3.5.1.  There was a compatibility break in 4.0 with Check_MK, an addition that's too valuable to lose.  It's since been resolved, I think, but I haven't gotten around to upgrading yet.  I will probably put it all together on a new CentOS 7 VM for maximum fun.

> I am under orders to use packages and not compile, a viewpoint I endorse.
> Are there other repos folk use?

I use the distro package manager to manage Nagios' dependencies, but I compile Nagios myself for maximum control.  It's really not that hard.  What I would like to do next is get the best of both worlds and learn how to make my own Nagios RPM.

> What about front ends, visualization etc.?

I use Check_MK.  It can replace Nagios' own configuration files, but I only use that part for several Linux hosts.  However, the LiveStatus pages is a Swiss Army knife compared to the stock Nagios pages (which are still available).  And it's FAST.

I don't use any Nagios-specific add-ons for Nagios' config files.  However, since they are still just text files, I use a programmer's text editor (SciTE), version control (git), and rsync to move config files from the dev Nagios host to the production Nagios host to the DR Nagios host.

Both Check_MK's LiveStatus pages and SciTE understand regex searches.  Unfortunately, they use different syntaxes, but once you get the hang of them they're invaluable.

> Any comments about FAN?

I don't know anything about FAN or any other Nagios configuration tools, but the moment you have to do something creative with your monitoring system will probably be the same moment you figure out it's not something automatic tools were programmed to cope with.  Might as well get comfy with the config files from the start.