Use Cases
Armada KV was designed around a single insight: most systems need fast local reads but only need to write centrally. Here are the scenarios where that model wins.
One source of truth, everywhere
Push application configuration from a central cluster to every edge location in real time. Services read config locally with sub-millisecond latency while the leader cluster handles all writes, keeping every region in sync automatically.
Ship safely at global scale
Store feature flag state in Armada and read it at the edge with microsecond latency. Roll out features progressively region by region using the hub-and-spoke topology — the leader controls the rollout, followers serve it instantly.
Global reads, central writes
Armada's follower clusters act as read replicas that are always current. Use them to serve routing tables, rate-limit policies, or any lookup data that must be fast everywhere but only changes occasionally.
Centrally managed, locally enforced
Distribute secrets metadata, API gateway policies, or ACL rules from a single secured leader cluster to all edge nodes. Armada's Raft-backed leader guarantees your writes are durable before any follower receives them.
Consistent state without the complexity
Use transactions to implement compare-and-swap counters, distributed leases, or lightweight coordination metadata. Armada gives you multi-key ACID transactions backed by Raft — no separate coordination service needed.
Always-on, even when regions fail
Follower clusters replicate the full dataset asynchronously. If the leader region goes down, followers keep serving reads. Promotion of a follower to leader restores full write capacity with a single operational step.
Spin up a single-node Armada cluster in minutes and start exploring the API.
Armada KV