reconcile clusteroperator status more frequently (#3006)

If status is modified, the operator must set status to proper values. This ensures that accidents cannot permanently reset status and gives a clear indication that the operator is "live". This came up as important when operators were NOT live, during cert rotations and we had no indication of problems.

Signed-off-by: Luis Sanchez <sanchezl@redhat.com>

pkg/lib/operatorstatus/status.go

@@ -228,7 +228,7 @@ func MonitorClusterStatus(name string, syncCh <-chan error, stopCh <-chan struct
  // if we've reported success, we can sleep longer, otherwise we want to keep watching for
  // successful
  if successfulSyncs > 0 {
- time.Sleep(5 * time.Minute)
+ time.Sleep(25 * time.Second)
  }
 
  }, 5*time.Second, stopCh)