Rancher Monitoring: Support decoupled prometheus stack

[richard-cox]

Is your feature request related to a problem? Please describe.

• The rancher monitoring chart installs prometheus, including grafana, rancher grafana dashboards, and services that expose those dashboards
• Given Decouple the Rancher monitoring stack while preserving Rancher UI dashboards ob-team-charts#221 the chart will no longer install prometheus+grafana

Describe the solution you'd like

• We need to ensure install and upgrade of this chart and the functionality it provides is least broken
• TBD Different chart?
  • old chart - install / upgrade - guide to new chart
  • new chart - cleaner install (no old chart upgrade)
• Install
  • Users will need to BYO prometheus stack and ensure certain configuration is supplied
    • Orbs + Docs task
  • Users will need to install the new rancher monitoring chart with ease
    • custom monitoring component used during helm install needs updating
    • link to docs explaining change (prom stack + grafana removed)
    • UI + Docs task
• Upgrade
  • Breaking change - process TBD
    • Migration Strategy?
  • Orbs + Docks task
• Integration
  • Granfana Dashboards are embedded in the following places in the cluster explorer
    • Cluster --> Dashboard --> cluster metrics
    • Cluster --> Dashboard --> kube metrics tab
    • Cluster --> Dashboard --> etcd metrics tab
    • Cluster --> Nodes --> click on a node --> Metrics tab
    • Cluster --> Workloads --> Deployment (any other than pod) --> click on a pod --> Metrics tab
    • Cluster --> Workloads --> Deployment (any other than pod) --> click on a pod --> Project Metrics tab
    • Cluster --> Workloads --> Pod --> click on a pod --> Metrics tab
    • Cluster --> Workloads --> Pod --> click on a pod --> Project Metrics tab
    • Cluster --> Monitoring --> Monitors --> Project Monitors tab --> click on one
      • TBD: Are there others in that Monitors tab group?
  • options like range and refresh should be wired in
  • links to grafana both in dashboards AND in the Monitoring home page
• Support matrix - prom/granfan to rancher
  • TBD: Product / Orbs team

Changes require co-ordinate with @manno and Orbs team

Additional context

• Mario's proof of concept changes master...manno:rancher-dashboard:configurable-monitoring-urls

[richard-cox]

@manno some questions...

• How are Monitoring --> Monitors --> Project Monitors configured and shown?
• What's the expectation around alerting, should that continue as before?

[github-actions]

🤖 Issue grooming questions

The issue provides solid context and a detailed solution outline, but several key items are marked TBD, making it difficult to define the scope of Dashboard changes. To groom this issue, please address the following:

1. Acceptance criteria for the Dashboard scope — Several items are TBD (upgrade process, support matrix, project monitors). Can you define, as a concrete checklist, the specific Dashboard changes that are in scope for this issue? For example: which embedded Grafana dashboard locations must render correctly, and which chart configuration UI changes (questions.yaml wiring, etc.) are required?
2. What does "done" look like from a Dashboard perspective? A verifiable checklist of UI behaviours would help engineers estimate and reviewers confirm completion.

This comment was generated automatically to help gather the information needed to groom this issue.

Generated by Daily Issue Grooming · ◷