Currently, the operator is running locally algonside of the tests, which maked some of the issues that only appear when running in cluster proper.

We should update the CI in such way this i tested closer to a production setup.

It is not clear we can use openshift/konflux/hypershift becuse this is an upstream project, but at least k3d or k3s or virtual cluster setup could be close enough to proper installation so that we would see the issues sooner.