Nagios Support Forum

Nagios® Core™ 3.2.0
ESXi 5.0


Hello Everybody,

I search for the best indicators that give me an hint that something is wrong with the ram and memory management on the Windows vm 's ( win7,xp,server2012).

I've read that the monitoring of the whole cpu and ram Parameters ( cpu usage, Pagefile, ram usage .etc) does not make much sense, due to the circumstances that the Windows System is using the Pagefile ( swapping) all the time
and should be.

Furthermore the cpu usage is also useless ,cause if the cpu has a large average it says nothing about Cpu Problem. Monitor the waiting processes like unix would do the Job much better ( are there any wmi or perfmons for that indicator available ?).


Thanks for helping !!!


regards skyerjoe

Perfmon has monitoring parameters under "system" "processor queue lenght". I think that's what your looking for. Allthough CPU usage does give you a hint of the workload of a individual VM, so it's perfectly fine to monitor that. When CPU hits as high as 80-100% over a longer period of time, something is wrong, except when you are running something like scanning software or other CPU intensive processing. As to RAM memory, you are basically looking for peak memory usage which is shown in the windows taskmanager. I dunno what kind of systems you are trying to monitor, since that really makes out, what you want to see. I wanna see different things on a terminal server then on a sequal sever or exchange.

But Im nog nagios expert, so can't tell you the commands to monitor these

Hello Ingmar

First of all Thanks for your answer

Do you have some hints to Monitor the ram usage in a proper way for me too ?


reagards skyerjoe

With most memory checks in windows you will get separate info on the page file usage, for instance with the standard NRPE or check_nt memory checks through NCPA or NSClient you should get the current usage, in both physical and page file. I believe you can also use performance counters as noted above.

I would monitor pages/s and amount of RAM usage. If that amount is low, and physical memory usage is low also, then there there is no problem with memory consumption on that system. It simply means, that the amount of requested pages that were not immidiatly available in RAM is low, so if you check then for disk read activity, there won't be alot of disk reads from your pagefile (I tend to setup up seperate pagefile disks on VM's). Even if RAM usage is high, but the amount of pages/s is low, your basically fine. It's when pages/s, combined with disk reads/s on the paging disk is high, you may wanna look into it. Cause that could cause performance issues.

So high RAM usage and high page reads per second would be a nice combination to monitor for I think. Correct me if I'm wrong

I completely agree with ingmar. The only addition I would have, is that if you wish, you could use the check_cluster service or bischeck to do some additional comparison of the usage between the two and statistical increases compared with previous data points. Bischeck in particular would give you a good resource to see if there are any recent abnormal spikes in usage.

Thanks guys for helping

best regards skyerjoe