For 4.6+

Support considerations for monitoring

Installing custom Prometheus instances in monitoring:

• You can install a custom Prometheus instance, the support is there.
• Remove the following from the "Support considerations for monitoring" section:
  "Installing custom Prometheus instances on OpenShift Container Platform."
• There are are considerations and limitations that need more explanation if the user wants to install their own version. Create a separate section to explain:
  • Prometheus Operator version needs to be the same (APIs and API groups need to be the same/on the same version)
  • If the user doesn't want to use a specific Prometheus Operator version, the user can install the default Prometheus Operator
  • With the default installed, other components are also installed
    • Add a note in the new section that recommends that the user installs the same Prometheus Operator version that is available in the default Monitoring stack if they want to install a custom instance. 

• What happens if there are two different versions of the Prometheus Operator and their associated CRDs?
  • The custom deployed operator might break if it is a different version than the default.
  • As long as CVO is running, it checks for the expected version of the CRD
  • The user needs to use the same version as what is shipped by the CMO, i.e. the default Prometheus Operator version
  • Include steps on how to check the logs and/or metrics for the current, default Prometheus Operator version in Understanding the monitoring stack.