We have a need to ignore namespaces, how can i cleanup this config for merge to main?
What happens is when you load configmaps for whatever reason all the leader elections update every X seconds and then kube-mgmt then bombards OPA with reloads. This has the effect of driving load thru the roof, see screenshot below. This is a 5 second trace and the heap is going crazy!!!
We have a need to ignore namespaces, how can i cleanup this config for merge to main?
What happens is when you load configmaps for whatever reason all the leader elections update every X seconds and then kube-mgmt then bombards OPA with reloads. This has the effect of driving load thru the roof, see screenshot below. This is a 5 second trace and the heap is going crazy!!!
CPU before/after