Regular update requests to calico-system namespace

[paulgmiller]

Tigera operator seems to send about 25 -50 update requests to apiserver for /api/v1/namespaces/calico-system' every five minutes.

Not really a problem for me but a customer complained this was noisy and crashing another security component that was monitoring namespaces. So I looked a little bit deer, got some errors around ippool to go away but the updates persisted. In the end looking at audit logs I could only find two diferences in requests and resposne for the namespace. The request did not include finalizders and hid a slightly differnt creation timestamp format.

This may be expected behaviour but thought I'd report in case it was unintentional.

[paulgmiller]

Actually that image was the diffence between requst and kubectl get Response from api serve did have managed fields too

[tmjd]

Is the namespace changing or is it just the requests themselves that are causing the problem even though they don't change the namespace?
I brought up an AKS cluster with calico, because that is what it looks like this cluster is you're using, and I watched the namespace and saw no changes after 30+ minutes.

[paulgmiller]

Namespace is not actually chaninging they have some other component (dynatrace) that is intercepting the requests and having a bad time.

[github-actions]

This issue is stale because it has been open for 30 days with no activity.