openskills.info
Vitess logoCourse Preview

Vitess

Vitess is a database clustering system that presents many MySQL instances as one logical database. It routes queries, manages shards and replicas, and supports operational changes as a MySQL deployment grows.

itCloud native tools and technologies

Vitess

Vitess is an open-source database clustering system for scaling MySQL horizontally. It sits between your application and a fleet of MySQL servers. Your application connects through the MySQL protocol, while Vitess routes each query to the right server.

This changes where scaling logic lives. Without a database-aware routing layer, an application may need to know which server owns each row. Vitess instead describes that placement in a VSchema and applies it through VTGate. You keep the relational model and much of the MySQL interface while adding a control layer for sharding, replication, failover, and data movement.

Vitess is not a replacement storage engine. MySQL still stores the rows, maintains indexes, executes local queries, and performs replication. Vitess coordinates multiple MySQL instances and gives applications one logical entry point.

The architecture at a glance

A request follows this path:

application
    ↓ MySQL protocol
VTGate
    ↓ routed request
VTTablet
    ↓ local MySQL connection
MySQL

VTGate is a stateless proxy. It parses SQL, consults the VSchema, chooses one or more tablets, and combines results when a query spans shards. Applications normally connect to VTGate instead of connecting to individual MySQL servers.

VTTablet runs beside each MySQL instance. It manages query execution, connection pooling, replication-related operations, and the tablet's role in the cluster.

MySQL remains the data store. A shard normally has one primary for writes and may have replicas for reads, backup work, and other operational tasks.

The Topology Service stores small amounts of cluster metadata and provides coordination. Vitess supports etcd, ZooKeeper, and Consul as topology backends. VTGate also maintains health information from tablets, so query routing does not require a topology lookup for every request.

Continue the course

This section is part of the paid course.

See pricing to subscribe, or log in if you already have access.

Where this skill leads

Relevant careers

See how this topic contributes to broader role-level skill maps.

Sources