ClickHouse Cluster

Configure your cluster topology with flexible sharding and replication. Each shard can have multiple replicas for high availability and data redundancy.

• Dynamic shard configuration: Set the number of shards based on your data distribution needs
• Multi-replica support: Each shard can have multiple replicas for fault tolerance
• Automatic data distribution: ClickHouse handles data distribution across shards automatically

ClickHouse Keeper provides distributed coordination for your cluster without external dependencies like ZooKeeper.

• Built-in coordination: No need for external ZooKeeper clusters
• Odd-numbered quorum: Deploy 3, 5, or 7 keeper nodes for optimal performance
• Automatic leader election: Handles failover automatically

Optimize costs with automatic data tiering between hot and cold storage. Simply define your cold storage location and ClickHouse handles the rest.

• Automatic tiering: Data moves to cold storage based on age or access patterns
• S3-compatible cold storage: Use any S3-compatible storage for cold data
• Transparent access: Query hot and cold data seamlessly—no application changes needed
• Cost optimization: Keep frequently accessed data on fast storage, archive the rest

Each ClickHouse node includes a restore_replica.sh script that can recover a failed replica by streaming metadata and access control data from a healthy peer in the same shard. The script automatically stops the local container, transfers the necessary files via SSH, sets proper ownership, creates a force_restore_data flag, and restarts ClickHouse. This allows you to quickly restore a replica without manual intervention or complex recovery procedures.