Keys, Namespaces, and SCAN

What you'll walk away with

Explain the core ideas behind Keys, Namespaces, and SCAN
Run the sample Redis command or code and verify its output
Apply the technique correctly to the Tutorial Platform and production scenarios

Build the mental model

A Redis key is a binary-safe string in a global namespace. A pattern such as `app:entity:id:purpose` clarifies ownership and debugging and can include environment or tenant boundaries. `KEYS *` scans the entire dataset at once and can block production. Incremental `SCAN` is safer, but it is not a snapshot and may observe duplicates or miss concurrent changes.

Connect it to a real scenario

Use keys such as `tutorial:42:card`, `session:<token>`, `rate:user:42:<window>`, and `leaderboard:weekly:2026-W35`. Inspect unknown keys with `TYPE`, `TTL`, `MEMORY USAGE`, and `OBJECT ENCODING`, then iterate in batches with `SCAN MATCH tutorial:* COUNT 100`.

Use a consistent namespace
Avoid KEYS on production request paths
Do not treat SCAN results as a snapshot

Try the working example

shell

SET tutorial:42:card '{"title":"Redis"}' EX 300
TYPE tutorial:42:card
TTL tutorial:42:card
MEMORY USAGE tutorial:42:card
SCAN 0 MATCH tutorial:* COUNT 100

You should see

You get the key type, remaining TTL, memory estimate, and cursor-based scan results.

5-minute try-it

Design a key convention separating development/production and two tenants, then create five example keys.

One important caution

Putting raw user input directly into keys can create unbounded cardinality, collisions, and sensitive-data leaks. Normalize or hash it and enforce limits.

Redis — Keyspace — Redis

Easy traps

Putting raw user input directly into keys can create unbounded cardinality, collisions, and sensitive-data leaks. Normalize or hash it and enforce limits.
Validate sample commands on a local or test instance with recoverable data before applying them to production Redis.

Exercise

Design a key convention separating development/production and two tenants, then create five example keys.

You'll know it worked when: You get the key type, remaining TTL, memory estimate, and cursor-based scan results.