Nagios Support Forum

First, I have to mention I am very new at using any products like this and also looking at perfmon or WMI.

I am currently evaluating xi and also looking at Zabbix.

We currently have a few alerting systems setup but I am looking to consolidate these with a more robust and useful system.

One of the main things I would like to get out of a new implementation is the ability to go back in history and look at what process caused a CPU or memory spike.

We have a few hundred servers running and we will get questions on occasion from the business owner as to what caused the server to run slow at 5pm last Thursday....or similar questions.

Can I look back historically and find this information out with Nagios? If so how can I setup this monitor to test?

Thank you

Nagios does save historical information for hosts and services, but the historical data will retain the host/service state, status text, and performance data. You could take a look at our online demo at demos.nagios.com for an idea as to what information is retained. Tracking the "cause" of host and service problems would have to be inferred from data coming from other checks.

So your saying that directly we cannot go back and see what process was consuming 60% of CPU.

Is there a way to monitor specific problem processes so we can track that for historic purposes and/or trending analysis?

Ok, I see what you're saying. Yes, this can be done. You could create a check (if it doesn't already exist) that monitors CPU and/or memory utilization of the specific process. I looked up on http://exchange.nagios.org , and it looks like there are quite a few plugins already made for this general category, you may need to sift through them to find exactly what you need, but if you want to monitor resource usage for specific processes, that can most certainly be done in Nagios.