Understanding Redis Default 16 Databases: Concepts, Configuration, and Cluster Limitations

In many projects Redis is used for caching, distributed locks, and message queues, and a newly installed Redis server creates 16 logical databases by default.

Each database is essentially a separate dictionary within the Redis instance, similar to having multiple databases in a relational DBMS, and they are identified only by numeric indexes.

The number of databases can be changed by editing the databases setting in redis.conf and restarting the server.

Clients connect to database 0 by default, but can switch to another database with the SELECT command, for example:

# 切库
redis> SELECT 1 # 默认0号db，切换为1号db
OK
redis [1] > GET username # 从1号库中获取 username 
(nil)

Because Redis does not support custom database names or per‑database passwords, databases should be treated as namespaces; a single Redis instance should serve one application, while different numeric databases can be used to separate environments such as production and testing.

In Redis Cluster mode only database 0 exists, so the SELECT command is unavailable; cluster nodes also impose limits on batch key operations, transactions, and Lua scripts, and the FLUSHALL command clears all databases at once.

Overall, the default 16 databases are configurable namespaces rather than independent stores, and for isolation it is recommended to run separate Redis instances per application, noting that clusters support only a single logical database.