Debezium

Debezium is an open-source platform for change data capture (CDC), enabling applications to react to every row-level change in a database in real-time. Originally developed by Gunnar Morling and now a Red Hat-sponsored project distributed under the Apache 2.0 license, Debezium has over 11,000 stars as of 2026. Debezium is built on top of Apache Kafka Connect and transforms existing databases into event streams, capturing every insert, update, and delete operation as structured event records. Key features include: change data capture for multiple databases (with connectors for MySQL via binlog, PostgreSQL via logical replication slots, MongoDB via change streams, SQL Server via Change Tracking, Oracle via LogMiner or XStream, Db2, Cassandra, and Vitess), Kafka Connect integration (running as a Kafka Connect source connector, distributing CDC events to Kafka topics with partitioning and fault tolerance), event structure (emitting structured events with before and after states, source metadata including transaction IDs, timestamps, and database positions), schema evolution (tracking and propagating DDL changes via the schema history topic, ensuring consumers understand schema changes over time), Debezium Server (a lightweight standalone runtime for streaming changes to non-Kafka destinations like Pulsar, Kinesis, Redis Stream, and Pravega), transformation Single Message Transforms (SMT for filtering, routing, and content-based event modification), snapshot mode (capturing the initial state of existing data before switching to streaming mode for ongoing changes), incremental snapshots (performing consistent snapshots without locking tables, using watermarks to interleave snapshot and streaming events), and monitoring via JMX metrics.