apache/kyuubi

Health v2 · maintenance only

Adoption not part of the score

2361 stars · 1017 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Apache Kyuubi is a distributed, multi-tenant Thrift JDBC/ODBC gateway that provides serverless SQL access to data warehouses and lakehouses. It manages pre-warmed Spark, Flink, and Trino engine instances so end users and BI tools can run SQL against large-scale data without managing compute themselves.

Use cases

• run interactive SQL queries on Spark from BI tools via JDBC
• provide a multi-tenant SQL gateway for a data lakehouse
• serve HiveServer2-compatible connections for big data analytics
• run large ETL batch jobs through a unified SQL interface
• isolate compute resources per user or tenant on a shared cluster
• query Iceberg, Hudi, Delta, and Hive tables from one endpoint

When to choose

• you need a serverless, multi-tenant SQL entry point over Spark/Flink/Trino
• you want HiveServer2 compatibility so existing JDBC/ODBC clients and BI tools work unchanged
• you require per-user engine isolation, high availability, and high client concurrency
• you want unified authentication and authorization across multiple query engines

When to avoid

• you need a lightweight embedded SQL engine rather than a distributed gateway service
• your workloads are small-scale single-user queries where a direct Spark or Trino connection suffices
• you need OLTP-style transactional queries against a relational database
• you cannot operate JVM-based server infrastructure or a Hadoop/lakehouse storage layer

Facets

service · maturity stable

database api-framework auth authorization scheduling big-data databases analytics microservices jvm cloud self-hosted spark-sql jdbc-odbc-server lakehouse hive serverless-sql multi-tenancy etl-gateway apache-software-foundation sql data-engineering kubernetes docker linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="apache/kyuubi")