The ClusterMonitoring CRD in config/v1alpha1 uses an incorrect singular resource path in its kubebuilder annotation, causing client-go generation issues and test failures. The annotation uses "clustermonitoring" (singular) instead of "clustermonitorings" (plural), violating Kubernetes API conventions.

OpenShift API repository - config/v1alpha1/types_cluster_monitoring.go

Always

1. Examine config/v1alpha1/types_cluster_monitoring.go kubebuilder annotation
2. Run make update-codegen-crds to generate CRD manifests
3. Run make integration to execute test suite
4. Observe test failures due to CRD name mismatches between generated manifests and test fixtures

- kubebuilder annotation uses: +kubebuilder:resource:path=clustermonitoring (singular)
- Integration tests fail with "missing CRD files" errors
- Generated CRD manifests use plural names but test files reference singular names
- Client-go generated code uses incorrect endpoint paths

- kubebuilder annotation should use: +kubebuilder:resource:path=clustermonitorings (plural)
- Integration tests should pass
- CRD manifests and test files should use consistent plural naming
- Client-go should generate correct plural endpoint paths following Kubernetes conventions

Fix requires both automated regeneration (CRDs, OpenAPI) and manual updates (test directory names, crdName fields in test fixtures). The test framework intentionally keeps test files semi-manual, so kubebuilder resource path changes require manual synchronization of test suite files.

Root cause: Lack of automated validation to ensure kubebuilder resource paths match test file crdName references.