How to visualise cpu usage out of 100%

[mitom (Tamas Millian)]

Hi,

While trialing the elastic stack I noticed that some of my servers report above 100% cpu usage. After looking around I concluded this is because the scale for some reason is not 0-100%, but 0-cpu_cores*100% (which honestly feels very wrong). What I am wondering about is what are the options when one actually wants to work on the assumption that 100% is the maximum an instance can use?

The interface does not give any help. I assume I could just put in something like system.cpu.total.pct/system.cpu.cores as field, but that's not happening. I notice there is the advanced options button, but that once again does not point me towards anything helpful (not even the help tooltip).

Is there documentation on what the advanced options are? Some sort of reference instead of a "here is 1 example of what you can do here"?
Is there a way to just build a simple timeline that scales the usage to be between 0 and 100 without having to pre-process the data to actually have that value?
Is there a way to change the metricbeat configuration to actually report between 0 and 100? What even is the point of reporting in an un-normalised range and calling it percentage?

Thank you,

Tamas

[lukas (Lukas Olson)]

You could create a scripted field from system.cpu.total.pct/system.cpu.cores. Have you created scripted fields before?

[mitom (Tamas Millian)]

Thank you, I was not aware of the option, but got it to work now.

As a follow-up question - which is probably not related to Kibana: there are a lot of warnings for the performance implications of scripted fields and as I understand it these get computed at search time. Is there a better or more efficient way to achieve this?