Applications may need a shared, memory-based store for cached values, counters or other data structures. Valkey provides that service, with choices for persistence, replication and deployment topology. It can be evaluated for workloads in the Redis ecosystem, but familiar protocols do not establish compatibility with every command, module or client behavior. First decide whether the data is disposable or needs durable recovery, then test the specific workload and restore path on the chosen version.
Valkey has an official Helm chart announced in January 2026. It is a candidate when choosing an in-memory data platform, but a chart's existence does not establish compatibility with every Redis workload, module or managed service.
Identify what the data means
A disposable cache, a durable queue and a primary data store have different loss and recovery tolerances. Inventory client commands, Lua scripts, modules, persistence settings, replication, TLS, authentication and failover behavior before selecting a migration method.
Prove recovery and client behavior
Test the chosen version with representative traffic and a restored data set. Include reconnects, failover, expiring keys and partial unavailability. A rollback after writes begin at the destination needs a data plan, not only the prior chart values.
The Redis and Valkey guide helps separate protocol compatibility, operating topology and licensing decisions. This entry makes no universal drop-in compatibility claim.
Sources & further reading
Spotted something that needs another look?
Help improve this page →