Thousands of "Skipping cache lookup" messages

[alexisph]

Hi.

We've deployed OADP operator on OCP v4.6 via this procedure: https://github.com/konveyor/oadp-operator#getting-started-with-basic-install-non---olm

We have observed an immense amount of logs being emitted by the oadp-operator pod:

  ...
{"level":"info","ts":1620604499.8003464,"logger":"proxy","msg":"Skipping cache lookup","resource":{"IsResourceRequest":false,"Path":"/apis","Verb":"get","APIPrefix":"","APIGroup":"","APIVersion":"","Namespace":"","Resource":"","Subresource":"","Name":"","Parts":null}}
...
--------------------------- Ansible Task StdOut -------------------------------
...
TASK [velero : Set up velero controller] ***************************************
...

and so on.

The result is that our Cluster Logging Elasticsearch instance is being filled completely in 2-3 days.

How do we disable this excessive logging? I have found that the ansible-operator base image used by the oadp-opearator is still on v0.17.0. There was a fix that was released a few months ago and is not included in that version. It is included in v1.5.0 and later.

Refs:
openshift/openshift-restclient-python#384
operator-framework/operator-sdk#3933
https://stackoverflow.com/questions/67415884/how-to-reduce-the-log-verbosity-of-ansible-based-kubernetes-operator-deployment
operator-framework/operator-sdk#4519
operator-framework/operator-sdk#4511

Edit:
We built the oadp-operator using the ansible-operator:v1.5.0 as base image and the excessive logging has stopped.

[dymurray]

Thank you for reporting this. Sounds like upgrading to a newer ansible-operator base image will help us here.