Product Features

Lossless Cluster Restart

Full cluster recovery from either planned shutdowns or cluster-wide crashes.

Minimizes process disruption due to planned or unplanned events

Hazelcast’s Lossless Recovery feature provides advanced fault-tolerance support for high-volume, mission-critical applications

Automatic Replication and Persistence of Critical Data

Data is continuously snapshotted and backed up to the disk, allowing the seamless recovery from the cluster-wide failures and shutdowns.

Uninterrupted Performance for Customer-Facing Streaming Data

Snapshotting, replication and persisting automatically occur in milliseconds, with essentially no noticeable effect on user experience.

Features

Replicated Storage

Jet stores data in multiple replicas (copies) across the cluster, which are recovered from the back-up if a node fails.

Persistent State

Continuously persists the states of cluster members on disk in a format designed for restart performance and to work in concert with SSDs.

Checkpointing

Consistent snapshots of each job are saved to storage, configured to be persistent with Hazelcast Hot Restart.
 

Rewindable

Rewindable sources are rewound using offsets saved in the snapshot (Kafka, Hazelcast IMap, Hazelcast ICache data sources support rewinding).