Restart cluster or a specified Pod

[MegaByte875]

Is your feature request related to a problem? Please describe.

The essence of Pod restart is to restart the service process, for example, when conventional methods can no longer solve the scenario of memory leakage, the only solution is to destroy the Pod to release the occupied resources and then rebuild it.

Cluster restart can be divided into two scenarios:

Restart of all Pods under the component

• Add a mark to trigger the component restart
• The controller recognizes and restarts each Pod during the reconcile process
• If the component needs to transfer state, it needs to clear the current node status before it can be rebuilt

Restart of a single Pod under the component

• Add a mark to trigger the Pod restart
• The controller checks whether the Pod needs to perform state transfer operations
• Perform the delete Pod operation, the StatefulSet controller will watch and create a new Pod

Describe the solution you'd like

Describe alternatives you've considered

Additional context

[MegaByte875]

#385