Although KOGITO-290 added the required annotations to the pod and service, this approach do not work on OpenShift 3.11/4.x that relies on Prometheus Operator. The discussion about this can be viewed in:
https://github.com/coreos/prometheus-operator/issues/1547

The Prometheus Operators maintainers do not agree with the annotations approach. Thus to have our services to be compatible with either a manually provisioned Prometheus in the cluster or managed by the Prometheus Operator we will have to:

1. Identify if the Prometheus Operator is available in the cluster
2. Create a ServiceMonitor resource in the namespace being owned by the Kogito Operator

This way we increase the compatibility with Prometheus and can expose our metrics in any scenario.

This "known issue" has been documented in our README.