Managed node group reports join failure while nodes already appear Ready in the cluster

The console still shows a failure, but kubectl shows Ready nodes, sending the team into the wrong troubleshooting path.