Towards https://github.com/giantswarm/giantswarm/issues/2993. I am still rabbit holing and figured the broken connections are also related to ingress-operator and endpoint-operator. I am going to move the endpoint resource to the kvm-operator, which we wanted to do anyway. There we have the control in one place and can cancel reconciliation and keep finalizers on demand. Endpoint-operator here is super behind and I am not going to update this project since we want to retire it soon. So here we exclude new pods from processing because we handle them in the kvm-operator instead.
Towards https://github.com/giantswarm/giantswarm/issues/2993. I am still rabbit holing and figured the broken connections are also related to ingress-operator and endpoint-operator. I am going to move the endpoint resource to the kvm-operator, which we wanted to do anyway. There we have the control in one place and can cancel reconciliation and keep finalizers on demand. Endpoint-operator here is super behind and I am not going to update this project since we want to retire it soon. So here we exclude new pods from processing because we handle them in the kvm-operator instead.