Fix EC2NodeClass readiness to not stuck in False

We observed some EC2NodeClasses stuck in Ready=False with message of `Failed to detect the cluster CIDR`. We believe this is caused by a bug in Karpenter, and this commit fixes that.

The previous commit adds a failing test to reproduce and the demonstrate the issue, and this commit fixes the implementation.

Author: mumoshu

pkg/controllers/nodeclass/readiness.go

@@ -46,6 +46,9 @@ func (n Readiness) Reconcile(ctx context.Context, nodeClass *v1.EC2NodeClass) (r
 			nodeClass.StatusConditions().SetFalse(status.ConditionReady, "NodeClassNotReady", "Failed to detect the cluster CIDR")
 			return reconcile.Result{}, fmt.Errorf("failed to detect the cluster CIDR, %w", err)
 		}
+		// Ensure the ready condition is set to true once the cluster CIDR is resolved,
+		// to not stuck in Failed state indefinitely.
+		nodeClass.StatusConditions().SetTrue(status.ConditionReady)
 	}
 	return reconcile.Result{}, nil
 }