Reference fixture · clickhouse

The fastest open-source OLAP database.

Sub-second analytical queries over petabytes. A columnar engine built for real-time analytics, observability, and the agentic workloads of the next decade — running on commodity hardware.

Run a query Get started Read the docs

Cluster Operational

-- 1.2B rows · 38ms
SELECT country,
       count(*) AS hits
FROM   events
WHERE  ts > '2026-05-01'
GROUP BY country
ORDER BY hits DESC
LIMIT  10;

Press ⌘ K to open the query palette.

100×

Faster than row-oriented stores on analytical workloads.

38ms

P95 latency on 1.2B-row aggregations, single replica.

2.2PB

Largest production cluster — and growing every quarter.

What this fixture exercises

Engineered for the workloads other databases can't keep up with.

Featured

Columnar at the core

A vectorized execution engine reads only the columns a query touches. Compression ratios of 10–100× shrink storage and accelerate scans.

Inspect tokens →

Real-time ingestion

Millions of rows per second per node. Kafka, S3, JDBC, HTTP — ingest from anywhere, query while writes are still in flight.

Read the spec →

SQL you already know

ANSI SQL plus a powerful expression library. Window functions, JSON, geospatial, arrays — every analytical primitive, native.

Browse functions →

Form components

Spin up a cluster in 30 seconds.

Inputs use the same charcoal-bordered geometry as the cards. Focus ring is neon volt — unified interactive feedback across the system.

Already running ClickHouse? Connect an existing cluster.