vitessio/vitess ? reverse-engineered prompt

Build me a cloud native database clustering system for MySQL that can scale horizontally with sharding.

I want the main idea to be that apps can keep talking to one database endpoint while the system handles spreading data across multiple MySQL servers behind the scenes. It should support splitting and merging shards, and those changes should be able to cut over cleanly without much downtime. Please make it feel like a real production ready open source project, with sensible docs, examples, tests, and the basic tooling needed to build and run it locally.

If you need to check current best practices or up to date docs while you work, go ahead and look them up online. Focus on the core database orchestration and scaling behavior first, then make sure the repo is organized well enough that someone could actually understand how to use it.