Data rebalancing needs to be initiated manually by the end-user. It depends on the specific use case and end-user environment if this makes sense. When end-users want to isolate new workloads and corresponding data on new nodes, data rebalancing is not used.