OLM caused kube-apiserver, and consequentially etcd, outage due to API requests overflood. 

4.13.34

$ zcat audit.log.tar.gz  | jq -Rr 'fromjson? | select(.requestReceivedTimestamp | contains("2024-03-11")) | (.responseStatus.code|tostring) + " " +  " [" + .user.username + "] " + (.verb|ascii_upcase) + " " + .requestURI'  | sort | uniq -c | sort -n  | tail -5 
36827 200  [system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount] LIST /apis/admissionregistration.k8s.io/v1/mutatingwebhookconfigurations?labelSelector=olm.owner%3Dcryostat-operator.v2.4.0-3%2Colm.owner.kind%3DClusterServiceVersion%2Colm.owner.namespace%3Dcryostat
36829 200  [system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount] LIST /apis/admissionregistration.k8s.io/v1/validatingwebhookconfigurations?labelSelector=olm.owner%3Dcryostat-operator.v2.4.0-3%2Colm.owner.kind%3DClusterServiceVersion%2Colm.owner.namespace%3Dcryostat
36884 200  [system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount] LIST /apis/operators.coreos.com/v1/namespaces/cryostat/operatorgroups
37052 404  [system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount] GET /apis/operators.coreos.com/v1alpha1/namespaces/cryostat/clusterserviceversions/cryostat-operator.v2.4.0
39703 201  [system:serviceaccount:openshift-apiserver:openshift-apiserver-sa] CREATE /apis/authorization.k8s.io/v1/subjectaccessreviews?timeout=10s

$ zcat audit.log.tar.gz  | jq -Rr 'fromjson? | select(.requestReceivedTimestamp | contains("2024-03-11") ) | select(.requestURI=="/apis/operators.coreos.com/v1alpha1/namespaces/cryostat/clusterserviceversions/cryostat-operator.v2.4.0")' | wc -l                                                   
 37052

// ** Showing the first and the last time
      the request was executed (doing this to get the time range) **

$ zcat audit.log.tar.gz  | jq -Rr 'fromjson? | select(.requestReceivedTimestamp | contains("2024-03-11") ) | select(.user.username=="system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount") | .requestReceivedTimestamp' -r | egrep -o "2024-03-11T.*:.*.:" | uniq | sed -e 1b -e '$!d'
2024-03-11T11:39:
2024-03-11T16:36:

OLM should pool these resources via API at more relaxed intervals to avoid throwing down kube-apiserver.