[eridius]

> For example, the maximum value for CPU utilization that you can have regardless of load is 100%.

I'm surprised it can't use load average to estimate the true resource demand.

[wahnfrieden]

You can, but it’s not visible to the hypervisor (it’s an OS concept) so you have to publish that metric from an agent on the machine. Then you can use it for autoscaling.

[viraptor]

But even then, "load" doesn't work well for all workloads. http://www.brendangregg.com/blog/2017-08-08/linux-load-avera...

The number of switching tasks may be very high for a number of reasons, including a very large number of threads which do very small chunk of work each and yield.