Debezium Blog

One question that we encountered recently is how to effectively integrate change data capture (CDC) with AI workloads — particularly for scenarios in which critical organizational knowledge is not publicly available. To help you to take advantage of your internal data, Debezium 3.1 introduces AI-focused features such as the Embeddings SMT and the Milvus sink, which you can combine to supply inputs to an LLM. You can read more about these enhancements in the Debezium 3.1 release notes.

Another release cadence done, and we’re pleased to announce the next preview release of Debezium is available, 3.2.0.Alpha1. This release is built on Kafka 4.0 with several breaking changes with many improvements and bugfixes.

Let’s take a moment and dive into all these changes.

It has only been three weeks since we released Debezium 3.1.0.Final, and we’re happy to report the first maintenance release has arrived, 3.1.1.Final. This release includes several critical performance improvements and a variety of bug fixes.

We’re writing to share an important update regarding the Debezium team. As part of a broader strategic move, the Debezium team, along with Red Hat’s Middleware and Integration Engineering and Products teams, will be transitioning to IBM in July 2025.

A year ago we began this incredible journey to create a modern approach to Change Data Capture. We had the desire to create a tool where you can focus on your data, defining how it flows from sources to destinations, with a pipeline-based approach. All this paired with a new and modern user interface to simplify interaction with it.

We named it Debezium Management Platform, or if you prefer, just Debezium Platform.

We are excited that Debezium 3.1 is the first official release of this years-long effort.