On a regular base there are huge spikes of uncaptured CPU time.
Why are they happening?
Check is the SCHEDULE and CSCHEDUL commands..
Several of these scheduled tasks run at 15 seconds, 30 seconds, 1 minute, 5 minute, or at 1, 4 or 12 hour intervals.
When all of the data collection occurs at times which coincide this could lead to a large 'spike' in the CPU.
When all of the tasks run simultaneously i.e. at a certain time the hit to the CPU becomes noticeable.
Release : 16.0
Component : SYSVIEW
Adjust the schedule tasks to spread out the intervals so that they rarely coincide to smooth out spikes.