Valkey 9.0 enhances in-memory databases with atomic slot migration, per-field hash expiry, and cluster support for better performance and scalability in real-time applications.

Valkey 9.0 has arrived, delivering major enhancements for in-memory database users. This update introduces atomic slot migration, per-field hash expiry, and cluster database support – key features that boost performance and scalability for real-time workloads.
Atomic slot migration allows entire data slots to move between cluster nodes atomically, eliminating the performance bottlenecks of key-by-key transfers. This is especially useful for in-memory database setups requiring minimal downtime. The new per-field hash expiry commands let developers set time-to-live on individual hash fields, enabling more granular data management without purging entire records. Additionally, numbered databases are now fully supported in cluster mode, improving data separation and making Valkey a stronger system benchmark candidate for complex deployments.
These improvements position Valkey as a compelling performance profiler companion and Redis alternative. The atomic migrations reduce cluster rebalancing time significantly, while per-field expiry optimizes memory usage for caching scenarios. For developers working with API client integrations or message queues, these features provide finer control over data lifecycle management.
Valkey 9.0 represents a significant step forward for open-source in-memory databases. With atomic slot migrations, granular hash expiry, and robust cluster database support, it addresses key pain points for scaling real-time applications. These features make Valkey an increasingly attractive option for developers seeking high-performance data storage solutions.
Atomic slot migration allows entire data slots to move between cluster nodes as a single operation, eliminating the performance issues of moving keys individually during cluster resharding.
Per-field hash expiry lets developers set individual time-to-live values on specific hash fields, enabling granular data management without deleting entire hash records when only some fields expire.
Cluster database support allows using numbered databases in cluster mode, improving data separation and multi-tenant isolation for complex deployments.
Atomic slot migration moves entire data slots between nodes as a single operation, reducing resharding time and eliminating performance degradation during cluster rebalancing.
Per-field hash expiry is useful for caching scenarios and API integrations where fine-grained control over data lifecycle is needed, optimizing memory usage without purging entire records.