Hi all, was curious if anyone knows of a method to retrieve CPU register values (or PECI data) from within ESXi via CLI or GUI? Under linux you can use a tool like rdmsr to accomplish this. I'm asking because I'd like to retrieve TEMPERATURE_TARGET, or thermal margin, etc. to determine if my hardware is allowing CPU's to operate outside of Intel's specifications.
Another solution is with Nagios (I've tried this and it's valid). It takes time to set it up but it's worth it.
Unfortunately two register values I need are not part of vsphere's performance metrics, so they can't be viewed using what you posted, nor would they be accessible to nagios. What vmware reports under hardware health for cpu temp is what the platform sends it, which is the Tcase temp value at the CPU heat spreader. That unfortunately does not tell me if the CPU is being allowed to reach its thermal limit and activating thermal control / performance reducing as a result. In my case I'm running Cisco UCS hardware that is being allowed to reach its thermal threshold constantly throughout the day, Cisco is telling me everything is great just ignore that because the CPU core temp is not being allowed to reach its limit, but I don't believe them so I'd prefer to be able to prove them wrong.
I understand, in PowerCLI I recommend you write directly to LucD who is a genius. If there is something he can do for your problem, he surely knows how to do it.
It depends on what the hardware sends. This is a bit outdated so you may need to test
Monitoring ESX hardware with PowerShell - VMware PowerCLI Blog
Re: Script to get serial number of all hosts
I haven't tested this my self but the articleshows using this to pull rpm metrics temperature metrics are probably in there somewhere too.
It might be better to post this again and post it in the powershell area, you have a good chance LucD will see it, but it helps if you at least have a good first attempt started.