forked from rook/rook
fullstackjam
f635d96ae9
This PR addresses ArgoCD drift detection issues by adding consistent metadata labels to RBAC resources in Rook Ceph Helm charts. The problem occurs because many RBAC resources lack metadata labels, causing ArgoCD to interpret them as 'labels: null' and report continuous drift. Changes made: - Added consistent labels to all RBAC templates in rook-ceph and library charts - Created _recommended-labels.tpl template for standardized labeling - Removed app.kubernetes.io/component label per consideRatio feedback to avoid scope creep - Suppressed Helm-specific labels (version, instance, managed-by, created-by, chart) from static RBAC files - Regenerated deploy/examples/common.yaml and deploy/examples/csi/nfs/rbac.yaml Labels added to RBAC resources: - operator: rook - storage-backend: ceph - app.kubernetes.io/name: rook-ceph - app.kubernetes.io/part-of: rook-ceph-operator Impact: - Resolves ArgoCD 'OutOfSync' issues for Rook RBAC resources - Improves resource identification and management - Provides consistent labeling across all Rook Helm charts - No functional changes to RBAC permissions - Focuses on essential labels to avoid over-engineering This addresses review feedback from travisn and consideRatio to maintain scope and ensure static RBAC files only contain meaningful labels. Signed-off-by: fullstackjam <fullstackjam@outlook.com>