Doing performance tuning on distributed systems is no joking matter. It’s much more complicated than on a single node server, and bottlenecks can pop up anywhere, from system resources in a single node or subcomponent, to cooperation between nodes, to even network bandwidth. Performance tuning is a practice that aims to find these bottlenecks and address them, in order to reveal more bottlenecks and address them as well, until the system reaches an optimal performance level.
In this article, I introduce you to TiDB, a distributed NewSQL database, and share some best practices on how to tune write operations in TiDB to achieve maximum performance.
