test(home): cover realProbeEnv discovery (live/degraded/no-release) via the cluster-load seam

[LukasWodka]

Summary

Fourth increment from the CLI test-coverage audit — the home-screen environment probe's discovery paths. realProbeEnv was 35%: it called cluster.Load / cluster.NewClientset directly, so a test could only reach the ownership-gate and load-failure returns — the actual release discovery (the core of the default bare-tracebloc screen) was dark.

Production change (seam — zero behavior change)

realProbeEnv now routes through the existing loadClusterFn / newClientsetFn seam instead of calling cluster.Load / cluster.NewClientset directly (both default to the real helpers, so production is unchanged). This lets a fake clientset drive discovery.

Test

• TestRealProbeEnv_Discovery (new) — drives the three discovered states with a fake clientset:
  • live (ready jobs-manager + a Ready node) → localLive, release name resolved, compute surfaced;
  • degraded (jobs-manager present, ReadyReplicas 0) → localDegraded;
  • no release (reachable, empty cluster) → localNoRelease.
  • The ownership-gate + load-failure paths stay covered by the existing TestRealProbeEnv_OwnershipGate.
• withClusterSeams now sets RestConfig — realProbeEnv does resolved.RestConfig.Timeout = …, so the seam's fake ResolvedConfig needed a non-nil RestConfig. Additive; its existing resolveClusterTarget users don't read it (all their tests still pass).

Not in this PR (follow-up)

runClusterInfo (cluster.go:134) and runClusterDoctor (doctor.go:99) still call cluster.Load directly. Their routing + post-discovery tests (which also need a MintIngestorToken reactor + printing assertions) are the next increment.

Test plan

• make ci green (build, -race, gofmt, lint, schema). New probe tests + the existing resolveClusterTarget tests (which share withClusterSeams) all pass.

🤖 Generated with Claude Code

---

Note

Low Risk
Test-only seam wiring with production defaults unchanged; no auth, data, or user-facing behavior changes.

Overview
Adds test coverage for the home screen’s realProbeEnvrelease discovery paths, which were previously hard to exercise because kubeconfig load went straight to cluster.Load / cluster.NewClientset.

realProbeEnv now calls the same loadClusterFn / newClientsetFn hooks as resolveClusterTarget (still defaulting to the real helpers in production). TestRealProbeEnv_Discovery drives a fake clientset through live (ready jobs-manager + node → localLive + compute), degraded (jobs-manager not ready → localDegraded), and no release (localNoRelease).

withClusterSeams now supplies a non-nil RestConfig on the fake ResolvedConfig so tests don’t panic when the probe sets RestConfig.Timeout.

^{Reviewed by Cursor Bugbot for commit 4ef436e. Bugbot is set up for automated code reviews on this repo. Configure here.}

[LukasWodka]

bugbot run

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 4ef436e. Configure here.}

[saadqbal]

LGTM 👍 Clean seam routing — and good that you set RestConfig: &rest.Config{} in withClusterSeams, since realProbeEnv now sets .Timeout on it through the seam. The live/degraded/no-release split is asserted right.