gracefulEvictionTasks is added to rb, but is never automatically removed, even if the cluster is ready

[CharlesQQ]

What happened:
Failover and GracefulEviction featureGate is open.
After the cluster becomes notready, the following content will be added to rb by the controller:

gracefulEvictionTasks:
  - fromCluster: xxx
    producer: TaintManager
    reason: TaintUntolerated
    replicas: 2

However, when the cluster is ready, the above content will not be removed from rb, affecting the scheduling of new expansion pods to the cluster.

What you expected to happen:
After the cluster is ready, gracefulEvictionTasks removes the content about the cluster from rb

How to reproduce it (as minimally and precisely as possible):

• Modify the ApiServer address of a member cluster to an unreachable address
• Wait for cluster notready
• Resize deployment replicas
• Wait for rb to add gracefulEvictionTasks
• Restore the ApiServer address of the cluster
• Wait for the cluster to be ready
• resize deployment replicas, whether the cluster has newly scheduled instances and the contents of gracefulEvictionTasks in rb

Anything else we need to know?:

Environment:

• Karmada version:
  v1.12
• kubectl-karmada or karmadactl version (the result of kubectl-karmada version or karmadactl version):
• Others:

[RainbowMango]

Can you share the PropagationPolicy of the Deployment?

The creation of a graceful eviction task indicates that the failover is about to process, was the failover successful?