Skip to content

Cluster references are hardcoded to one deployment #4

Description

@stef9github

The console resolves the VSS stack through src/lib/cluster-refs.ts, which hardcodes namespaces, deployment names, ConfigMap keys and service URLs for the deployment it was built against. A cluster that names things differently — a different VSS profile, a different namespace layout, an object store other than ARTESCA — will have pages report a component as absent when it is running.

Nothing fails loudly, which is the awkward part: the affected panels render their degraded state, so the symptom reads as "that component is down" rather than "the console looked in the wrong place".

What would resolve this

Turn the cluster references into configuration, with the current values supplied as one example rather than as the definition. A deployment should be able to state its own namespaces and service names without a rebuild.

Notes for anyone picking this up

  • src/lib/cluster-refs.ts is the single source of truth the rest of the tree reads; the pages themselves do not hardcode names.
  • docs/console-design.md states the operator-facing intent of each page, which is the right reference for deciding what each lookup is actually for.
  • The reachability strip on / and the /diagnostics page are where a wrong reference becomes visible first, so they are a reasonable place to start.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions