Split from #39 (frontend debt umbrella). This is the "re-enable skipped suites" item, scoped and root-caused.
The 10 skipped tests
K8sClusterList.test.tsx — 3 (shows Active status badge, shows Unreachable status badge, uses connected status as primary truth)Fleet.test.tsx — 3 (prioritizes connected access status, keeps connected primary truth, shows diagnostic probe status secondarily)k8s-explorer.integration.test.tsx — 4 (loading skeleton, empty state, disables generic ingress creation on OpenShift, mutation-boundary notice)
Root cause (verified empirically)
The K8sClusterList / Fleet reachability skips are not "just needs an MSW mock" as their inline TODOs say. The reachability display migrated: ClusterStatusBadge (ui/ClusterStatusBadge.tsx) now reads from the SSE-driven reachability registry via useTargetConnectivity(targetType, targetId) — the cluster probe for direct/operator clusters, the SSH probe for jumphost clusters. It no longer consumes the batch GET /api/k8s/clusters/connectivity result these tests mock. Un-skipping them today fails with Unable to find text: Reachable/Unreachable, because the mocked batch probe never reaches the badge.
So re-enabling them means rewriting the tests to drive the reachability registry (seed useTargetConnectivity / its provider), not mocking the old endpoint — and updating the stale label assertions (connected probe → the badge's current label). The k8s-explorer four have separate causes (loading/empty-state timing, OpenShift-context gating) and should each be checked individually.
Acceptance
Each suite's it.skip removed and passing against current component behavior, or — if a test encodes behavior that intentionally no longer exists — deleted with a one-line note rather than left skipped. No it.skip with a stale "needs a mock" TODO left in these three files.
Split from #39 (frontend debt umbrella). This is the "re-enable skipped suites" item, scoped and root-caused.
The 10 skipped tests
K8sClusterList.test.tsx— 3 (shows Active status badge,shows Unreachable status badge,uses connected status as primary truth)Fleet.test.tsx— 3 (prioritizes connected access status,keeps connected primary truth,shows diagnostic probe status secondarily)k8s-explorer.integration.test.tsx— 4 (loading skeleton,empty state,disables generic ingress creation on OpenShift,mutation-boundary notice)Root cause (verified empirically)
The
K8sClusterList/Fleetreachability skips are not "just needs an MSW mock" as their inline TODOs say. The reachability display migrated:ClusterStatusBadge(ui/ClusterStatusBadge.tsx) now reads from the SSE-driven reachability registry viauseTargetConnectivity(targetType, targetId)— the cluster probe for direct/operator clusters, the SSH probe for jumphost clusters. It no longer consumes the batchGET /api/k8s/clusters/connectivityresult these tests mock. Un-skipping them today fails withUnable to find text: Reachable/Unreachable, because the mocked batch probe never reaches the badge.So re-enabling them means rewriting the tests to drive the reachability registry (seed
useTargetConnectivity/ its provider), not mocking the old endpoint — and updating the stale label assertions (connectedprobe → the badge's current label). Thek8s-explorerfour have separate causes (loading/empty-state timing, OpenShift-context gating) and should each be checked individually.Acceptance
Each suite's
it.skipremoved and passing against current component behavior, or — if a test encodes behavior that intentionally no longer exists — deleted with a one-line note rather than left skipped. Noit.skipwith a stale "needs a mock" TODO left in these three files.