During a baseline audit of the remediation operators in the openshift-workload-availability namespace, a discrepancy was found in the priorityClassName configuration. While the Fence Agents Remediation (FAR) and Node Healthcheck (NHC) controllers are protected by system-cluster-critical, the SBR (sbd-operator) controller-manager is running with no defined priority class (defaulting to 0).

In a resource-constrained environment, the SBR controller could be preempted/evicted to make room for application workloads, effectively breaking the node fencing mechanism when it is needed most.

Environment:

• Namespace: openshift-workload-availability

• Operator: Storage-Based-Remediation (SBD)

Observed Priority Mapping:

Steps to Reproduce:

1. Inspect the pods in the remediation namespace:

   oc get pods -n openshift-workload-availability -o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.priorityClassName}{"\n"}{end}' | grep sbd-operator 

1. Observe that the result for the controller-manager is empty.

Expected Result: The sbd-operator-controller-manager should have priorityClassName: system-cluster-critical to ensure parity with other Medik8s/remediation components and guarantee availability during node pressure events.

Actual Result: The pod runs with default priority (0), making it a candidate for eviction by the kube-scheduler during heavy cluster load.