The inplaceupgrader controller in the HyperShift Configuration Operator (HCCO) logs an error when it encounters a degraded node during an upgrade, but it does not specify which node is degraded. Additionally, the NodePool CR status remains "Healthy" (e.g., AllNodesHealthy: True) even when the reconciler is blocked by a degraded node, leading to a "silent" upgrade failure from a user's perspective.

OCP management Version: 4.18.27
ACM version: 2.13.5
Platform: Hosted Control Planes (HCP) Agent
Version: OpenShift 4.17.x 
Targeting 4.17.37)
Upgrade Strategy: InPlace

    NA

During a customer HCP agent upgrade to 4.17.37, the upgrade stalled. The NodePool CR indicated that all nodes were updated and healthy, yet the HCCO logs showed a reconciliation error.     

HCCO Log Output:

{
  "level":"error",
  "ts":"2026-02-02T20:39:43Z",
  "msg":"Reconciler error",
  "controller":"inplaceupgrader",
  "object":{"name":"infra01-nodepool","namespace":"infra01-infra01"},
  "error":"degraded node found, cannot progress in-place upgrade. Degraded reason: disk validation failed: expected target osImageURL ... have ...",
  ...
}

The log error "degraded node found" does not mention the Node Name, making it extremely difficult for admins to troubleshoot clusters with high node counts (43 nodes in this case).

The NodePool status conditions (AllNodesHealthy, AllMachinesReady) reported True with the message "All is well," despite the upgrade being blocked by a validation failure.

The HCCO logs should explicitly state the name of the degraded node (e.g., "error": "degraded node [node-name] found...").

The NodePool status conditions should reflect that the upgrade is stalled and identify the failing node/machine in the message field.

    Required data will be uploaded and shared over the comment box