Is this a bug report or feature request?

• Bug Report

Deviation from expected behavior:
After active manager reelected, rook removes monitoring labels from mgr ServiceMonitor.

Expected behavior:
After active manager reelected, mgr ServiceMonitor should keep all labels.

How to reproduce it (minimal and precise):

1. Create Ceph Cluster with enabled monitoring and at least 2 managers:

spec:
  labels:
    monitoring:
      my-label: abc
  mgr:
    allowMultiplePerNode: true
    count: 2
    modules:
      - enabled: true
        name: rook
  monitoring:
    enabled: true

2. Check that monitoring labels were propagated to mgr ServiceMonitor:

% kubectl -n rook-ceph get servicemonitor rook-ceph-mgr -o yaml | yq .metadata.labels
team: rook
my-label: abc

4. Downscale active mgr (find one with metadata.label.mgr_role == active) and wait until remaining mgr will be labeled as active:

% kubectl -n rook-ceph scale deployment/rook-ceph-mgr-a --replicas=0

6. Check that now mgr ServiceMonitor is missing monitoring labels:

% kubectl -n rook-ceph get servicemonitor rook-ceph-mgr -o yaml | yq .metadata.labels
team: rook