[Improvement] Rethink the config of 'rss.server.write.slow.threshold'

### Code of Conduct

- [X] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)

### Search before asking

- [X] I have searched in the [issues](https://github.com/apache/incubator-uniffle/issues?q=is%3Aissue) and found no similar issues.

### What would you like to be improved?

The 'rss.server.write.slow.threshold' config option is intent to indicate the writing slow operations by summarizing writing time. But unfortunately, for huge event(single flush data event size reaches 10g) and different storage type(HDFS), it's not convincing.

I prefer to choosing the writing speed to indicate writing slow operation.

### How should we improve?

_No response_

### Are you willing to submit PR?

- [ ] Yes I am willing to submit a PR!

No files, tests, or entry points are named. Start by locating the rss.server.write.slow.threshold configuration and the code that evaluates write duration, then examine how large flush events and HDFS storage affect that measurement. Define the intended speed-based behavior and its completed configuration and test coverage before implementation.

Written by the indexing model from the issue text.