Skip to main content

Scaling for Rediss

Scaling Redis means spreading data and load across several nodes. Redis Cluster does this by splitting the keyspace into 16384 hash slots assigned to primary nodes, each with replicas for failover. Clients route commands to the right node, and multi-key operations need keys that share a slot.

Stage: Reliability · Part of the Redis roadmap · Last reviewed 19 November 2026

What to learn in Scaling

Scaling interview questions

Why do multi-key commands sometimes fail in Redis Cluster?

Each key belongs to a hash slot hosted on one node, and a command touching keys in different slots cannot run on a single node, so it returns a CROSSSLOT error. Hash tags, such as {user:1}:cart, force related keys into the same slot.

What does a MOVED response mean?

It tells the client that the slot for the requested key lives on another node, giving that node's address. Cluster-aware clients update their slot map and retry there. It is normal after resharding or a failover, and a good client handles it without surfacing an error.

Where Scaling fits in the Redis roadmap

Scaling belongs to the Reliability stage (intermediate level), which takes about 9 hours. Persistence and scale.

Back to the full Redis roadmap