YAML indentation was 2 levels deeper than siblings (14 spaces vs 12),
causing the key to be nested under httpGet instead of readinessProbe.
kubectl apply would silently accept this mis-structured manifest but
the probe configuration would be ignored at runtime.
PRI-907
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Add pre-deployment node/namespace/resource diagnostics and wrap
kubectl apply in explicit error handling with cluster state dump on
failure. This gives us actionable output in the GitHub Actions logs
when the Deploy E2E step fails, instead of a silent exit code.
PRI-956
Co-Authored-By: Paperclip <noreply@paperclip.ing>
The rook plugin E2E workflow was calling the reusable plugin-e2e workflow
without the required plugin-name input. This caused the ConfigMap naming
and mount path to fall back to the headlamp-kube-vip defaults, breaking
E2E runs for the rook plugin.
Fix: pass plugin-name: rook to the reusable workflow.
The 'Storage Classes' link is nested under the Rook sidebar button, not
at the top level. Expand the Rook section before asserting visibility.
Also uses /rook/i case-insensitive regex and waits for the plugins list
to render before searching.
Chris Farhood