When layered products want to integrate with OpenShift platform monitoring, they must create Prometheus resources in a namespace watched by the OCP platform Prometheus instance. Such namespaces must be labelled as 
openshift.io/cluster-monitoring: 'true'. One example of such a namespace is openshift-monitoring.
 
The current best practice for layered products is to label their own operand namespace with the above label, and thus store their Prometheus resources isolated from other OCP components. This workflow creates problems for layered products such as ACS, which allow installation into existing customer managed namespaces. The use of Helm, which does not manage namespaces and their labels natively, complicates the namespace creation additionally. It is therefore technically not always feasible to go the isolated namespace route. As a crutch, Prometheus resources are stored in openshift-monitoring instead, which creates friction with both OCP core maintenance and SRE operations (see linked ticket).
 
This feature request is essentially the ask for a dedicated namespace in OCP, in which layered products are free to Prometheus resources such that they are both

• isolated from customer namespaces.
• isolated from core OCP namespaces.

The namespace should always exist and carry the OCP platform monitoring label. Customers should be discouraged to create their own resources in the namespace via support voidance similar to openshift-monitoring.