Skip to main content

Quick start

Run Baleia on your machine with Docker, connect the bundled Polaris catalog and run a first query. It takes a few minutes, most of them pulling images.

Start the stack​

You need Docker with Compose and make.

Start Baleia and its dependencies
git clone https://github.com/maltzsama/baleia.git
cd baleia
make up

make up starts Baleia (API, worker and migrations) with everything it needs:

  • PostgreSQL for Baleia's own data.
  • Temporal for long-running work.
  • RustFS as S3-compatible storage for query results and the lake.
  • Apache Polaris as an Iceberg catalog, seeded with the catalog baleia_polaris and the table analytics.events.
  • Trino with the catalog-store plugin.
  • Mailpit to catch e-mails.

Optional profiles add more sources:

  • docker compose --profile sources up -d adds MySQL, MariaDB, SQL Server, Oracle, ClickHouse, Lakekeeper, Hive Metastore, Unity and a mock AWS.
  • --profile ldap adds a directory for LDAP sign-in.
ServiceAddress
Baleiahttp://localhost:8080
Temporal UIhttp://localhost:8233
Polaris consolehttp://localhost:4000
Mailpit (e-mails)http://localhost:8025

Create the admin account​

Open http://localhost:8080. On the first visit a setup wizard creates the admin account; after that you sign in normally.

The stack registers the bundled Trino as the engine trino-default on start, from TRINO_URL.

Connect the bundled catalog​

In Connections, click New connection. The wizard walks through the steps:

  1. Pick Apache Polaris and fill in:
    • URL http://polaris:8181, as the containers reach it;
    • client ID admin and client secret admin.
  2. Test connection.
  3. Add catalog with the warehouse baleia_polaris. Baleia creates the matching catalog in Trino: that is a binding.
  4. Grant access to your workspace, so it can query the catalog.

The connection types and their fields are listed in Connections.

Run a query​

Open SQL, pick the catalog in the target selector and run:

A first query
SELECT * FROM analytics.events LIMIT 10;

Run with ⌘/Ctrl + Enter. The workbench guide covers scripts, results, exports and saved queries.

Stop​

Stop
make down      # stop, keeping the data
make down-v    # stop and delete the volumes

Next​