GitHub uses Vitess to manage 150 terabytes of data. In this talk, hear Arthur Schreiber talk about how Vitess is used at GitHub to handle a 750k QPS database workload.
Vitess 25 is coming in October with several improvements to switchovers:
- Emergency failovers no longer wait on lagging replicas
- New primaries are chosen more consistently
- Operators have more control over split-brain recovery
Read more:
Horizontal sharding solves a whole category of problems. Vitess gives you faster recovery times, lower write contention, and more so you just keep scaling. 📈