I noticed that when Audit Scanner starts to run all request load is targetted into a single policy server pod, instead of being spread out throughout all available replicas, making the pod struggle to respond to all Audit scanner requests and making the auditing process take much longer. It does seem that Audit scanner also has very little concurrency when validating resources. I tried it on a large cluster (thousands of namespaces) and at the moment is taking around 2 days to finish.
Expected Behavior
Audit Scanner should take advantage of all replicas available from policy server to spread the request load
Is there an existing issue for this?
Current Behavior
I noticed that when Audit Scanner starts to run all request load is targetted into a single policy server pod, instead of being spread out throughout all available replicas, making the pod struggle to respond to all Audit scanner requests and making the auditing process take much longer. It does seem that Audit scanner also has very little concurrency when validating resources. I tried it on a large cluster (thousands of namespaces) and at the moment is taking around 2 days to finish.
Expected Behavior
Audit Scanner should take advantage of all replicas available from policy server to spread the request load
Steps To Reproduce
No response
Environment
Anything else?
No response