The `podSelector` of the `openshift-adp-velero-metrics-svc` service [1] matches the velero pod and the `repo-maintain-job pods` created from the maintenance job [2].

Therefore the service will trigger the TargetDown false alerts since the job is in `Completed` state instead of  `Ready`.  

[1]- https://github.com/openshift/oadp-operator/blob/45861be2086dbd36c53c1a0067eeb1e74021a886/docs/oadp_monitoring.md#create-oadp-service-monitor

 [2]- https://github.com/openshift/oadp-operator/blob/master/docs/kopia_troubleshooting.md#kopia-repository-maintenance