Looking for an explanation of CPU usage units in the webUI... usage vs time vs load

[kernull]

I find myself going down a rabbit hole when googling this, trying to understand the difference between usage, load, and cpu time...

I notice that I have some VMs whose summary shows and avg or ~3.5 but if I console in (windows vm) it shows ~ 10% usage. Likewise with some other VMs (linux too).

Whats the easiest way to mental-math my way to a metric of utilization percentage?

thanks for reading!

 

[dcsapak]

maybe when you show some examples, we can try to explain? it's a very broad topic (as you already noticed) so there isn't really some short summary that entails everything to know
one thing that could be interesting to know is that the hypervisor has a different view of what the vm does (from outside) than what the vm can see (from inside), so a difference
in these two views is normal and expected

 

[kernull]

@dcsapak solid points, sorry for being vague.

So to be more specific, if the CPU Usage metric shown in the graph is unit-less, what does it represent? (What is a value of 4.5 mean on a VM that has 2 vCPUs?)

thanks!

 

[Dunuin]

The "CPU usage" graphs are showing the CPU usage in percent. So a 4.5 in the "CPU usage" graph of the VMs summary page will mean 4.5%. In case that VM got 2 vCPUs that would be a load of 0.09 (because CPU usage of 100% would be a load 2.0).