OCP/Telco Definition of Done
Epic Template descriptions and documentation.

Epic Goal

• Restrict Platform Monitoring usage to core functionality: Telemeter, Basic Console
• Move the same functionalities of CMO into operator installable through OLM (installs UWM like a stack) with separate versioning to Openshift

Why is this important?

• We have to add more advanced features into our collection stack (agents, scalability, tenancy, security). Given our experience, we have to break out of core Openshift versioning that brings our velocity down.
• Performance of core monitoring component. We don't have enough control for the current platform monitoring data we ingest, due to various customers using it, so we cannot slim the core components down.

Scenarios

• Customers want to run Openshift with large cluster, scale out Prometheus, use soft multi-tenancy or run efficient single node clusters
• We want to reduce amount of metrics PM ingest due to scalability
• We want to add more complexity to CMO without extra baggage of having to support previous versions for long duration. (Decouple + separate versioning).
• Customer wants to get platform metrics related to their applications into their observability system (UWM or PR or anything else).

Acceptance Criteria

• CMO only installs Platform Monitoring
• CMO does not allow ad-hoc querying/dashboard
• CMO does not allow different remote writes other than Telemeter
• UWM-like stack is installable through OLM and has all features that current PM has
• UWM, or any other Prometheus compatible collection can scrape (access ServiceMonitors) platform metrics too (openshift namespaces too). PM is no longer a strict interface to the platform metrics.

Previous Work (Optional):

• Design doc: https://docs.google.com/document/d/18SAkrbBbibgAE-eS5BZZ1UU-I-Kc2CS_Rl0o0b35SN4/edit