On a OCP 4.18 cluster with 67 nodes and 1157 namespaces, observability-operator pod does not run stable as it crashes continously due to OOM a few seconds after being scheduled.

Looks like that the operator is running slightly uses more memory as its limit of 512Mi right after the start of the pod, compare attached screenshot.

Issue is reported in https://access.redhat.com/support/cases/#/case/04368491