Description of problem:

The ServiceMonitor used for the Loki Operator matches more than one service, which causes Prometheus to get more than one target for the Loki Operator. This results in duplicate metrics from the single operator instance.

Version-Release number of selected component (if applicable):

How reproducible:

Install Loki Operator to a cluster and enable "Operator recommended cluster monitoring on this Namespace" during installation, then look for the number of active targets in the namespace.

Steps to Reproduce:

1. Install Loki Operator with monitoring enabled
2. Query for targets: up{namespace="openshift-operators-redhat"}

Actual results:

Result should only show the Loki Operator once, but the query returns more than one series

Expected results:

Loki Operator should only show up as a single metrics target.

Additional info:

This has gone unnoticed for a while, but I am considering it a higher priority for the 5.9 release because we have added a new warning alert that will show up in a lot of customer clusters which would be complicated by the alert showing up multiple times.