vocabulary:
  name: Vitess Vocabulary
  description: >-
    Vocabulary for Vitess, the CNCF graduated database clustering system for
    horizontal scaling of MySQL. Covers sharding, topology, replication, and
    cluster management terminology.
  version: "1.0"
  created: "2026-05-03"
  tags:
    - Cloud Native
    - CNCF
    - Database
    - Distributed Systems
    - MySQL
    - Sharding

terms:
  - term: Vitess
    definition: >-
      A CNCF graduated open source database clustering system for horizontal scaling
      of MySQL. Vitess provides MySQL protocol compatibility, automated resharding,
      query routing through VTGate, and connection pooling, making it suitable for
      large-scale MySQL deployments on Kubernetes.
    tags:
      - Core
      - Platform

  - term: Keyspace
    definition: >-
      A logical database in Vitess that maps to one or more MySQL databases depending
      on whether it is sharded. An unsharded keyspace maps to a single MySQL database;
      a sharded keyspace is distributed across multiple MySQL database instances.
    tags:
      - Core
      - Database Structure

  - term: Shard
    definition: >-
      A horizontal partition of data within a Vitess keyspace. Each shard owns a
      contiguous range of rows identified by a key range (e.g., -80 or 80-). Each
      shard has one primary tablet and zero or more replica tablets.
    tags:
      - Core
      - Sharding

  - term: VTTablet
    definition: >-
      A server process that runs next to a MySQL instance, converting gRPC requests
      from VTGate into MySQL queries. VTTablet manages connection pooling, query
      execution, health checks, and replication. Each tablet has a type: PRIMARY,
      REPLICA, or RDONLY.
    tags:
      - Architecture
      - Core

  - term: VTGate
    definition: >-
      A stateless proxy that serves as the entry point for client connections to a
      Vitess cluster. VTGate routes MySQL queries to the appropriate VTTablet instances
      based on the routing schema (VSchema), handles scatter queries, and manages
      cross-shard transactions.
    tags:
      - Architecture
      - Core

  - term: VTAdmin
    definition: >-
      The administrative REST API and web application for managing Vitess clusters.
      VTAdmin provides a unified interface for inspecting topology, tablets, keyspaces,
      schemas, and VReplication workflows across one or more clusters.
    tags:
      - Administration
      - API

  - term: VSchema
    definition: >-
      The VTGate routing schema that defines how queries are routed across shards.
      The VSchema specifies vindexes (sharding functions), vindexed tables, lookup
      vindexes, and sequence tables for each keyspace.
    tags:
      - Core
      - Routing

  - term: Vindex
    definition: >-
      A sharding function in Vitess that maps a column value to a keyspace ID used to
      identify which shard owns a row. Primary vindexes are used for sharding; lookup
      vindexes are secondary indexes that map non-shard-key columns back to shard keys.
    tags:
      - Sharding
      - Core

  - term: VReplication
    definition: >-
      The Vitess data movement and transformation framework used to power operations
      like MoveTables, Reshard, Materialize, and CreateLookupVindex. VReplication
      reads the MySQL binary log and applies changes to target tablets, enabling
      zero-downtime data migrations.
    tags:
      - Replication
      - Data Migration

  - term: MoveTables
    definition: >-
      A VReplication workflow that moves tables from one keyspace to another without
      downtime. MoveTables is used to split a monolithic keyspace into separate
      keyspaces or to re-shard data for better distribution.
    tags:
      - Data Migration
      - VReplication

  - term: Reshard
    definition: >-
      A VReplication workflow that changes the number of shards in a keyspace by
      splitting or merging shards. Reshard redistributes data across more shards to
      handle growth or fewer shards to reduce overhead.
    tags:
      - Sharding
      - VReplication

  - term: Cell
    definition: >-
      An availability zone or isolated failure domain within a Vitess deployment.
      Cells are used to organize tablet placement and control cross-cell replication.
      A typical deployment has cells in different data centers or availability zones
      for high availability.
    tags:
      - Architecture
      - Topology

  - term: Topology Service
    definition: >-
      A consistent distributed key-value store (typically etcd or ZooKeeper) used by
      Vitess to store cluster topology information including keyspace, shard, and tablet
      metadata. VTGate and VTTablet query the topology service to discover cluster state.
    tags:
      - Architecture
      - Infrastructure

  - term: Primary Tablet
    definition: >-
      The tablet that currently handles all write traffic for a given shard. In a
      Vitess cluster, exactly one tablet per shard is the primary at any time.
      Failover to a new primary is handled automatically through planned or emergency
      reparenting operations.
    tags:
      - Replication
      - High Availability

  - term: vtctld
    definition: >-
      The Vitess topology management daemon that exposes a gRPC and HTTP API for
      administrative operations including creating keyspaces, managing shards,
      executing planned reparents, and emergency reparents.
    tags:
      - Administration
      - Architecture