CPU Monitoring frequency

This support forum board is for support questions relating to Nagios XI, our flagship commercial network monitoring solution.
Locked
polarbear1
Posts: 73
Joined: Mon Apr 13, 2015 4:26 pm

CPU Monitoring frequency

Post by polarbear1 »

Greetings.

Suppose I have 30 16-core Linux boxes where I am interested in monitoring CPU usage at the core level with as frequent interval as XI will allow.

Can someone tell me what the recommendation is for what this frequency is? I understand there are factors at play such as my network config, the horsepower of my XI box, and how much overhead I want to sacrifice to monitor this, so a best practice ballpark is as good an answer as I'm looking for.

Thanks.
rkennedy
Posts: 6579
Joined: Mon Oct 05, 2015 11:45 am

Re: CPU Monitoring frequency

Post by rkennedy »

It really depends on what you deem necessary, but I would think 3 minutes should suffice. If things are critical, then you could go lower.

You could always go as low as a one minute, which is the lowest I'd recommend. It'll take time for SNMP to communicate, which I've seen take up to 90 seconds before. (depends if they're LAN or WAN, and a couple other factors) With this lower number, it'll allow you to see the spikes that may be occurring, where as 3 minutes wouldn't pick up on them as much.
Former Nagios Employee
Locked