Openshift default monitoring stack and Service Mesh Operator

Question

I am using Openshift v4.6 and I use the default monitoring stack with user defined projects enabled. Openshift's documentation clearly states:

Custom Prometheus instances and the Prometheus Operator installed through Operator Lifecycle Manager (OLM) can cause issues with user-defined workload monitoring if it is enabled. Custom Prometheus instances are not supported in OpenShift Container Platform.

I now want to install Openshift's Service Mesh operator which for monitoring the mesh stuff wants to install a new Prometheus and Grafana instance. Doesn't this go against the above suggestion? I couldn't find anything in Openshift's documentation (or anywhere else) that could clarify this conflicting information. I actually moved forward and installed the operator with Prometheus and Grafana enabled and so far it doesn't seem to "cause issues" but I don't even know what issues to look for. Anyone with any experience on this?

Answer 1

Well, having two operators overwriting the same APIS can be an issue.

Custom Prometheus instances and the Prometheus Operator installed through Operator Lifecycle Manager (OLM) can cause issues with user-defined workload monitoring if it is enabled. Custom Prometheus instances are not supported in OpenShift Container Platform.

The potential CRD overwrites/incompatibility with a different version is the reason why the issue is a warning.

There are limitations to how K8s works w/ already registered APIs, you should deploy the same Operator version in order to avoid issues.

Does that make more sense?

Answer 2

I have installed Service Mesh on an environment and deployed several apps and did not find the need to reinstall Prometheus or Grafana. While using the default monitoring stack that comes with the OpenShift cluster, I was able to visualize the performance of the pods through Grafana without any issues. Hope this helps!