Skip to main content

ClickHouse databases

You’ll create a single-node ClickHouse database with durable disk / storage and HTTP plus native endpoints in the dashboard.

Unsure which path to use? Start at Choose a database path. This page is the Databases path.

Before you start​

  • Connected cluster and an environment
  • Permission to create a database on your role (owners and Full access members have it)
  • The Altinity ClickHouse operator installed on the cluster (system charts). Existing clusters need system charts re-run once.
  • Free disk for the size you pick. Prefer Ceph storage for production.

Limits (this release)​

  • One replica only — no shards, no ClickHouse Keeper, no ReplicatedMergeTree
  • Public access is HTTP 8123 (NodePort or ingress). Native protocol 9000 stays inside the cluster
  • Backups are volume snapshots, not ClickHouse-native freeze/backup

Create the database​

  1. Open Databases → New Database → ClickHouse.
  2. Select the environment.
  3. Set name, storage size, and CPU / memory (give ClickHouse several GiB of memory).
  4. Create and wait until status is healthy.

Connect an app​

  1. Open the database detail page and copy:
    • HTTP {name} service on port 8123 (SQL over HTTP, health /ping)
    • Native {name} service on port 9000 (official clients, JDBC)
    • Username default and the generated password
  2. Put them in the app’s variables or a shared variables group.
  3. Restart the app if needed, then confirm in logs.

In-cluster hosts look like clickhouse-{app-name}.{environment}.svc.

You’re done when​

  • Database status is healthy.
  • An app in the same cluster connects on HTTP or native protocol with the shown credentials.
  • Data survives an instance restart.

Turn on backups before production data.