Copied form mail conversation

2. Cluster Monitoring Configuration File Management

Prometheus requires both a monitoring system (typically installed with the cluster) and a specific configuration file. I observed that our operator attempts to create the cluster-monitoring-config ConfigMap (the configuration file) within the openshift-monitoring namespace if it doesn't exist. This approach introduces several issues:

a) Reference Prohibition: The operator attempts to set a reference to our webserver, but this type of reference is not allowed in OpenShift (at least in recent versions I've tested). This results in an immediate error logged by our operator.

b) Persistence Issue: If the ConfigMap is created without the problematic reference, it persists on the cluster even after the webserver and the operator are deleted/uninstalled.

c) Uninstallation Impact: Deleting this configuration upon operator uninstallation could potentially disrupt other applications/operators that users have installed after the JWS operator.

Proposal: From my perspective, the JWS operator should not be responsible for managing the cluster monitoring configuration file. Cluster monitoring configuration should be an administrative responsibility handled at the cluster level.