Nagios Support Forum

Hi,

We are currently trying to setup up ESXI host monitoring through our vCenter server in Nagios and are using the check_vmware_api.pl plugin. I have gone through the list of available counters in the API reference guide online and see that some of the metrics that are being requested for monitoring are not listed. We are looking to monitor cpu io wait for the ESXI host level as well as the host level BIOS information. I was hoping someone would have an idea about how to collect these metrics either through the plugins already available or any other method.

Thank you.

I don't know about io wait specifically, but have you reviewed the box293_check_vmware plugin on the exchange?
https://exchange.nagios.org/directory/P ... re/details

It's pretty comprehensive in what it provides as well as it's documentation.

Yes, I've already been through the documentation on that one and am currently using it in our development environment for a few different checks, but unfortunately I didn't find the metrics we are looking for there either. Thanks for the suggestion though.

I can't find much within the Nagios community that addresses iowait.

You might check your hypervisors compatibility with the Converter SDK as a potential framework, but some performance issues can occur and rapidly overload the Nagios machine when utilizing pretty much any of the VMWare SDKs.

I went through the API reference for the Converter SDK you linked to and it looks nearly identical to the vSphere SDK as far as available counters are concerned. For BIOS information, I looked into check the check_esxi_hardware.py plugin, but it seems to me that would require logging into the esxi host directly to get that data, and we are trying to avoid that as we have numerous hosts and want to get the data directly from vCenter. Am I correct in assuming at this point there is no available plugin for these specific metrics and that getting data from them would require a different approach?

If there aren't any sort of channels available from the various VMWare SDKs and APIs (without logging in directly), you're probably correct.

I have incredibly little experience with ESX, but if there's a way to evaluate iowait via the CLI you could examine it with a client-side script and ship the results to Nagios XI via NRDP.