impact on network load

[MichielvM]

Hi, I have a fairly simple question (I think), but finding an answer is not so simple..

I've been asked to provide information on how monitoring with Nagios/NSClient++ impacts network traffic/load. They want to know packet sizes sent from Nagios to Clients and back.
In this case we're looking at a maximum of 4 clients with a basic monitoring profile (i.e. default Nagiosxi's Windows server wizard)
Can anyone put a number on this?

[mp4783]

There's no way to put a precise number on this but you can use either a packet sniffer like WireShark or (on a Linux host at least) tcpdump to dump all traffic to/from the Nagios XI host to the client. You simply need to dump a single polling interval to one of your hosts and then total up the size of the packets.

What I can tell you anecdotally is that a standard set of active checks does not put much load on the network. Other than protocol and SSL overhead, the Nagios XI server just asks the monitored host to execute a plugin. The plugin then returns the output and an exit code. In most cases, this output will be relatively small.

To be honest, with only 4 clients and a standard monitoring setup, your network should barely even know Nagios is monitoring those servers.

[jdalrymple]

I profiled this once before. I wish I could find it but am unable to.

If memory serves me correctly though I found that an entire check_nrpe -c check_disk to a Linux server was like 7K of total traffic, aggregated, both directions. It's wildly negligible. Mind you by default the payload of an NRPE output is set to a maximum of 512 bytes.

If you can think of good search terms, search the forum for my posts, it's in there somewhere

[MichielvM]

Well, I already figured the load to be negligible. I think I can work with the info above.
Thanks!

[rkennedy]

Sounds good. I will close this out now as it's resolved, but feel free to open another thread if you need more assistance.