Distributed Key-Value Store for Kubernetes
A distributed, eventually consistent KV store with a hub-and-spoke replication model. Built for Kubernetes, optimized for high read throughput and global data distribution.
Armada brings together the best of distributed systems and Kubernetes-native operations.
Distribute data globally from a single core cluster to multiple edge clusters. Armada handles replication automatically—no cross-cluster complexity.
Manage and monitor Armada just like any other Kubernetes deployment. First-class support via the official Armada Helm Chart.
Built for read-heavy workloads. Serve sub-millisecond reads at the edge, making it ideal for configuration and feature data distribution.
Powered by the Raft consensus algorithm with data redundancy. Armada continues serving reads during network partitions and node outages.
Config distribution, feature flags, edge caching — see where Armada fits in your stack.
Explore use cases →Hub-and-spoke topology, multi-group Raft, MVCC versioning — understand how Armada works.
See how it works →Official Go and Java clients, community tools, and the gRPC API surface.
Browse the ecosystem →
Armada KV