What is AetherLake?

AetherLake is a batteries-included, Kubernetes-native Data Lakehouse that brings together best-in-class open-source tools into a single, cohesive platform. Instead of spending weeks gluing together storage, compute, catalog, orchestration, and security layers — deploy everything in minutes.

Key principles

• 🏗️ Modular — Enable or disable any component via a single toggle
• 🔐 Secure by default — One Keycloak login for every UI (native OIDC or an oauth2-proxy gate), RBAC realm roles, random per-install secrets, TLS via cert-manager
• 🌊 Streaming included — Kafka (Strimzi, KRaft) and Flink SQL jobs are part of the platform, with topics queryable through Trino
• 📦 Cloud-native — Helm charts, Kubernetes operators, and S3-compatible storage
• 🎛️ Unified control — Web-based Control Panel to manage the entire platform
• 🌐 Multi-language — Control Panel supports English and Turkish (extensible)

Core Technologies

AetherLake integrates several major open-source projects:

• Storage: MinIO (S3 compatible, operator-managed tenant)
• Table Format: Apache Iceberg
• Catalog: Apache Polaris (REST Catalog with S3 credential vending)
• Compute: Trino & Apache Spark
• Streaming: Apache Kafka (Strimzi) & Apache Flink (SQL jobs)
• Orchestration: Apache Airflow
• Analytics / BI: Apache Superset
• Vector DB: Milvus
• Identity: Keycloak (SSO for every service UI)

What you can do with it

• Query everything with SQL — Iceberg tables, Kafka topics, and external databases from one Trino endpoint or the Control Panel SQL IDE.
• Stream into the lakehouse — Flink SQL jobs read/write Kafka topics from the browser; produce from outside the cluster over an authenticated (TLS + SCRAM-SHA-512) listener.
• Explore the catalog — browse Iceberg schemas, tables, partitions and snapshots; manage Polaris namespaces and Trino catalogs from the UI.
• Operate the cluster — pod overview, live logs, Kubernetes events and CPU/RAM metrics without leaving the Control Panel.
• Automate with agents — the bundled MCP server exposes platform status, Trino queries, Polaris catalogs and Airflow DAGs to AI assistants.