Description of problem:
With a lot of customers going cloud/virtual it would be nice if Cockpit could show not just CPU load graphs but also have an option to show CPU stolen time used by other VMs on the host. When a customer is experiencing performance problems but showing negligible CPU load, it's nice to point out how they're getting robbed behind the scenes by other VMs on the host.
Version-Release number of selected component (if applicable):
0.93 F23 (current as of filing)
Steps to Reproduce:
1. Start Cockpit on Fedora.
2. Log into console and view a system's stats. CPU/Mem/Disk/Net graphs show.
CPU load shows but not all stats.
Would love a check box for extra CPU stats such as stolen time, per-core time, etc.
I know it may be an overflow of data to track stats for 16+ CPU cores but primarily it would be nice to track stolen CPU time for cloud or virtual deployments. Call me an old fashioned bare metal guy but customers should understand the performance implications of virtual deployments vs bare metal and overloaded virtual hosts. If nobody has done any work on something like this I will see if I can draft a feature poc on my own.
See See http://blog.scoutapp.com/articles/2013/07/25/understanding-cpu-steal-time-when-should-you-be-worried