Description of problem:
In a multinode OCP 4.8.35, on the nodes that PAO is installed, stalld process seems to have high CPU usage.

Version-Release number of selected component (if applicable):
Stalld version 1.13

How reproducible: often, according to the partner he observes that very often.

Steps to Reproduce:
Unclear how to reproduce it in a lab.

Actual results:
See every other second in top output that stalld takes from 30% to 70%

Expected results:
Stalld never be a process with high cpu usage.

Additional info:
This is a standard kernel, PAO reserves 8 cpus for the system and the rest for the workload. There is one at least DPDK pull mode process which keep one core at 100%, apart from that system is idle in overall. We have access on that system so we can easily check more info if needed.

Thank you