• Share this post Keep up with us Summary Zerobus Ingest is now Generally Available (GA), providing a fully managed, serverless service that streams data directly into Delta tables-eliminating the need for intermediate message buses like Kafka • Teams can achieve sub-5-second latency while supporting thousands of concurrent clients and delivering up to 100 MB/sec per connection for over 10 GB/sec of aggregate throughput to a single table • This GA launch includes production-ready gRPC API and REST APIs (Beta), plus SDKs for Python, Java, Rust, Go, and TypeScript, enabling integration from any application • Zerobus Ingest is now Generally Available (GA), providing a fully managed, serverless service that streams data directly into Delta tables-eliminating the need for intermediate message buses like Kafka • Teams can achieve sub-5-second latency while supporting thousands of concurrent clients and delivering up to 100 MB/sec per connection for over 10 GB/sec of aggregate throughput to a single table • This GA launch includes production-ready gRPC API and REST APIs (Beta), plus SDKs for Python, Java, Rust, Go, and TypeScript, enabling integration from any application
Article Summaries:
- Databricks has announced the general availability of Zerobus Ingest, a serverless data‑ingestion service that streams events directly into governed Delta tables within a Lakehouse. Designed to replace traditional message‑bus architectures such as Kafka, Zerobus Ingest offers a single‑sink, managed Databricks endpoint that eliminates intermediate layers, reduces operational overhead, and cuts costs. The service supports thousands of concurrent connections and delivers over 10 GB/s of aggregate throughput with sub‑5‑second latency. By writing data straight into Unity Catalog‑managed tables, it provides automatic lineage tracking and fine‑grained access control, streamlining governance and simplifying the overall streaming stack.
- Databricks has announced the general availability of Zerobus Ingest, a fully managed, serverless service that streams data directly into governed Delta tables. The new tool replaces traditional message buses such as Kafka by using a single‑sink architecture that pushes events straight to a Lakehouse, cutting out intermediate layers and reducing tool sprawl. Zerobus claims to support thousands of concurrent connections with aggregate throughput over 10 GB/s and sub‑5‑second latency to a table. By eliminating the compute and storage costs of a message bus and simplifying governance through Unity Catalog, the service aims to lower operational overhead and compliance risk while delivering high‑performance ingestion.
- As organizations scale real-time operational intelligence, traditional streaming architectures have become costly bottlenecks. Managing message buses like Kafka and handling schema registries and connector frameworks creates a significant “complexity tax” that diverts high-value engineering resources from strategic business initiatives. Meanwhile, duplicate storage inflates cloud bills and multi-hop architectures delay critical insights. Finally, data in transit often sits outside centralized governance frameworks, creating compliance risks and lineage blind spots. Today, we’re excited to anno
Sources:
- https://www.databricks.com/blog/announcing-general-availability-zerobus-ingest-part-lakeflow-connect
- https://www.databricks.com/blog/announcing-general-availability-zerobus-ingest (Latest source article published: 2026-02-23 16:00 UTC)