Add a timeout= option for the runSync()?

### Feature Request / Improvement

It would be worth adding a **timeout** option for the main syncing function to allow the system to return back to normal to investigate the problem without entering into resource leak problem linked to the execution or a costly cloud execution without full knowledge of the required time.
It is as basically saying I expect this sync to complete in less than 30 minutes, if not than it is not worth it before some data are changed before resyncing.

This feature can be valuable especially for full-sync of large tables.

May require the use of `https://docs.oracle.com/javase/8/docs/api/java/util/concurrent/CompletableFuture.html` in _XTableSync_ class' _runSync_())

[### Are you willing to submit PR?

- [x] Yes I am willing to submit a PR!](https://github.com/apache/incubator-xtable/pull/823)

### Code of Conduct

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

Start with the XTableSync.runSync() entry point and review the linked pull request 823 alongside Java's CompletableFuture documentation. Determine how the timeout should affect long-running full syncs, and verify that an expired sync returns control without leaving execution or resource leaks.

Written by the indexing model from the issue text.