AetherLakeOpen-Source Data Lakehouse

Storage, catalog, query, streaming, BI and identity — deploy the whole platform on Kubernetes with a single helm install.

🏗️ Modular Design

Enable or disable any component (Trino, Spark, Airflow, Kafka, Flink, MinIO…) via a single Helm value toggle.

One Keycloak login for every UI — native OIDC where it exists, an oauth2-proxy SSO gate where it doesn't — plus random per-install secrets.

Apache Kafka (Strimzi, KRaft) with Flink SQL jobs, and topics queryable as SQL tables through the Trino kafka catalog.

Iceberg tables on S3-compatible MinIO, Apache Polaris REST catalog with credential vending, federated SQL via Trino.

Manage the platform, write Flink SQL, browse Kafka topics, run Trino queries and watch pod logs from the Next.js Control Panel (EN/TR).

A built-in MCP server lets AI assistants check status, query Trino and manage catalogs and pipelines.