ClickHouse Ecosystem

ClickHouse is the columnar analytics database route when the task is about OLAP queries, event analytics, high-volume logs, or analytical SQL that should not be treated like transactional Postgres. Source: Ecosystem Capability Coverage, 2026-06-27; Skill Resolver, 2026-06-27

When To Use This

Use ClickHouse when workloads are append-heavy, analytical, time-series-like, or aggregation-first. Do not route ClickHouse tasks through generic Postgres habits without checking engine, partitioning, and query-shape assumptions. Source: Skill Resolver, 2026-06-27

Decision Matrix

Core Model

ClickHouse is optimized around analytical scans, compression, partitioning, ordering keys, and batched ingestion. The design question is usually "what query pattern must be cheap?" rather than "what entity relation is normalized?" Source: Ecosystem Capability Coverage, 2026-06-27

Routing Summary

Preferred order: ClickHouse MCP for schema/query inspection, clickhouse CLI for local or remote SQL, and ClickHouse best-practices skill before production query changes. Source: Skill Resolver, 2026-06-27; config/capability-harvest-seeds.json

Failure Modes

• Porting OLTP schema habits into analytical tables.
• Ignoring ordering keys and partitioning.
• Running unbounded exploratory queries against production data.
• Treating ClickHouse as a replacement for app transaction state.

Verification Path

Use clickhouse client --query 'SELECT 1' or MCP readback to confirm connectivity. For performance-sensitive work, show the bounded sample, query plan, or row-count guard used before full execution. Source: config/capability-harvest-seeds.json; Skill Resolver, 2026-06-27

---

Timeline

• 2026-06-27 | Canonical ecosystem page created so ClickHouse has database-pack placement, routing, CLI evidence, and a verification path. Source: User request, 2026-06-27; Ecosystem Capability Coverage