Single node
One host, one connection. Lightweight, zero topology probing.
- Direct queries against
system.* - No cluster overhead
- HTTP or HTTPS
Auto-detects topology — single node, self-hosted cluster, or ClickHouse Cloud. Monitor queries, kill runaway loads, manage access, and wake idle services.
Embedded UI kit with mock data across all three topology modes. Kill a query, drop a user, switch warehouses, toggle the theme.
Set the env vars and Chadmin figures out the rest. Same dashboard, mode-aware SQL underneath.
One host, one connection. Lightweight, zero topology probing.
system.*Failover across hosts; per-replica memory and state in one view.
clusterAllReplicas() queriesAPI-driven service discovery. Wake idle services on demand without polling them awake.
Built for the things you actually do at 3am when a query is eating the cluster.
30-second polling with pause-on-hidden — your idle Cloud services stay idle.
Per-replica KILL QUERY with confirmation. No SQL editor needed.
Browse system.users / grants / role_grants; drop accounts in a click.
Cloud-aware: control plane for state, data-plane ping only when you ask.
nginx + php-fpm + React, baked. Multi-arch: linux/amd64 and linux/arm64.
That ClickHouse-yellow accent, in both moods. Persists across reloads.
Pick a registry, paste the run command, open http://localhost:8080.
# Self-hosted ClickHouse — single host or cluster docker run -d --name chadmin -p 8080:80 \ -e CLICKHOUSE_HOST=clickhouse.example.com \ -e CLICKHOUSE_PORT=8123 \ -e CLICKHOUSE_USERNAME=default \ -e CLICKHOUSE_PASSWORD=secret \ bun4uk/chadmin:latest
Full env-var reference is in .env.example.
Available tags: latest, semver (M.m.p), M.m, M.
Need help? Open an issue.