What to monitor via SNMP

[shodg001]

Hello,

I wanted to get an idea of what things people are monitoring via SNMP. I'm getting ready to write some checks and as I am going through the MiB tree I found myself wondering what really is essential. For our purposes it will primarily be switches, routers, ASGs, and such. What specific things are you monitoring on your snmp enabled devices? Appreciate any input!

Regards,

Shawn

[slansing]

One place you may want to look for tips would be the exchange site, this is where community members upload their custom plugins and addons for the Nagios community:

http://exchange.nagios.org/

As far as some basic switch checking goes I typically see others monitoring port status "up/down" in other words "is something plugged into this port?" Performance data is another big one that people pull with SNMP, most of the time via rrdcached.

You should generally be able to search for your specific hardware on the exchange and come up with some plugins that people found useful.

[shodg001]

Hello,

To be honest I am not a big fan of using other people's plugins. I often find I have to dig through about 10 shyte ones to find one that works as advertised. In the time it takes to do that, I could already set up the snmp checks and move on the something else. So, anyone else want to add some input?

[sreinhardt]

This is a pretty open ended question being that what people consider important to monitor can widely vary. Personally on a basic level I look for:

network gear: interface status(up\down), port speed, and if available I like to verify the mac addrs that are in use on each port. I also have traps setup to tell me if the mac on a port changes, as the port then also goes administratively down(sticky macs or some such thing in cisco lingo).

Servers: I generally look for cpu usage, memory usage, disk usage if available, maybe total number of processes or if services are started. Some people will also look for smart statuses, raid status, and other things.

I realize that you may not want to use others checks, but at least looking at them or the most commonly used ones would give a good idea of some things you want to try. As you might expect, this is highly device, vendor, and version specific.

[shodg001]

Thanks for the reply. Yep I wanted this to be a wide open question. It's a "what do you do", so I can possible build a best practice for my company. I am interested in here EVERYTHING:)

[shodg001]

So here is a direct snmp question. I hope someone out there can help me puzzle this out. I am working hard on understanding MiBs and OiDs. One example that works, but I don't understand is the HP switch CPU usage.
The OID is .1.3.6.1.4.1.11.2.14.11.5.1.9.6.1.0. It works fine. But when I look at the MiB at http://www.oidview.com/mibs/11/STATISTICS-MIB.html they do not have it listed. And when I snmpwalk .1.3.6.1.4.1.11.2.14.11.5.1.9.6.1 (no 0 on the end) it does not show up at all. Can anyone account for this? I also don't often see a trailing 0 on the end of OIDs. Opinions, comments?

[sreinhardt]

That is entirely up to your vendors specification and configuration. As for the snmpwalk, often they index from 1 not 0. The mib, only thought I would have is being out dated.

With that said, I don't want to push you away from the forums at all, however it would be much appreciated if we could stay on nagios specific topics and not generalizations of other systems. By all means if you are having issues monitoring a specific OID or with a command we would love to hear it though.

[shodg001]

Sure. I would like to keep this thread open to hear what others are monitoring with Nagios via snmp.

[tmcdonald]

We generally try not to keep things open for hypotheticals. We get shocked if things go unresponded to for too long :D

If you have a specific question we'd be happy to help, otherwise I'm going to lock the thread.