The standard for distributed coordination
Naming, configuration management, synchronization, and group services — battle-tested coordination primitives in a simple, highly reliable interface.
Naming, configuration management, synchronization, and group services — battle-tested coordination primitives in a simple, highly reliable interface.
A proven coordination kernel trusted by Kafka, Hadoop, HBase, and many other distributed systems.
In-memory messaging with sub-millisecond latency; handles tens of thousands of requests per second.
A clean hierarchical namespace modeled after a file system makes coordination logic easy to reason about.
Every update is stamped with a monotonically increasing zxid, giving clients a total ordering guarantee.
Replicated across an ensemble of servers; survives node failures automatically with no data loss.
Common patterns and distributed systems problems where ZooKeeper excels.
Elect a single master among distributed workers reliably using ephemeral znodes and watches.
Implement mutexes and read/write locks across cluster nodes without a single point of failure.
Register and look up live service instances dynamically as nodes join or leave the cluster.
Propagate config changes to all nodes instantly with watches — no polling required.
Track which nodes are alive in real time using ephemeral nodes and group membership recipes.
Coordinate phased computations so all workers start and finish a phase together.
ZooKeeper is a top-level Apache project with an active community of users and contributors. Join discussions, read the documentation, and help shape the roadmap.
Up and running with ZooKeeper in a few simple steps.