Primary key table support standby replica

### Search before asking

- [x] I searched in the [issues](https://github.com/apache/fluss/issues) and found nothing similar.

### Motivation

Primary key table support standby replica.

During testing, I found that the primary reason for the slow recovery of PrimaryKey tables is the slow download of snapshot files, which accounts for about 90% of the total recovery time, while the apply cdc log phase takes up only a small fraction.

Therefore, could we introduce a standby replica that continuously synchronizes the latest snapshots? During recovery, we could restore directly from this standby replica.

This approach might even eliminate the need for rebalancing of PrimaryKey table.

### Solution

_No response_

### Anything else?

_No response_

### Willingness to contribute

- [ ] I'm willing to submit a PR!

Start by investigating PrimaryKey table recovery, snapshot-file downloads, CDC log application, and rebalancing in the Fluss codebase. Define how a standby replica would continuously synchronize snapshots and how recovery would restore from it; done means the design addresses recovery speed and the proposed reduction in rebalancing.

Written by the indexing model from the issue text.