cluster-operations

What problem does it solve?

Managing self-hosted ClickHouse clusters requires manual, error-prone work for distributed DDL propagation, node scaling, replication recovery, and infrastructure migrations, leading to downtime and data inconsistency if executed incorrectly.

Core Features & Use Cases

• Distributed DDL & Table Management: Automate schema changes across all shards with ON CLUSTER commands, configure distributed table sharding and read routing for optimal query performance.
• Node Lifecycle Operations: Safely add or remove cluster nodes, handle automatic data sync for replicated tables, and drain replication queues before decommissioning nodes to avoid data loss.
• Cluster Recovery & Migration: Restart failed replicas, reshard data with minimal disruption, and migrate from ZooKeeper to ClickHouse Keeper without extended downtime. Use case: For example, when scaling a ClickHouse cluster to handle a 2x increase in query traffic, use this skill to add new nodes, rebalance distributed table shards, and verify replication health without interrupting active workloads.

Quick Start

Use the cluster-operations skill to add a new ClickHouse node to your existing cluster and rebalance distributed table shards to handle increased query load.