Cluster Awareness

Valkey GLIDE was designed to handle the complexities of Valkey (and Redis) clusters automatically. It abstracts away slot management, node discovery, and failover logic so your application interacts with the cluster as if it were a single endpoint.

For information on Valkey Cluster, see this page.

Topology Aware

Requiring only one initial node within a cluster, GLIDE can use it to discover out the full topology and hash slot ownership. This allows GLIDE to calculates the destination of every command before sending it, ensuring they go directly to the correct node rather than bouncing around the cluster. If the cluster topology changes, either due to scaling or failover, GLIDE will automatically update its internal mapping.

Hash Slot

Valkey has a total of 16,384 fixed slots from which it splits across nodes in a cluster. Each key are then assigned to a slot for storage. GLIDE’s mapping of the cluster allows it to send them to the appropriate node.

Multi-Key Commands Handling

Valkey (and Redis) clusters support commands that operate on multiple keys which maybe spread across different hash slot. GLIDE cluster awareness allows it to efficiently handles these commands, ensuring correct executions across the cluster.

To learn more about multi-key commands, see this section.