New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
Every dashboard your team checks is already a few hours behind by the time anyone looks at it.
What you'll have at the end
A running Debezium connector that streams a table's inserts, updates, and deletes into a Kafka topic, confirmed live by a real update and a real delete.
You need
A PostgreSQL database you can configure yourself, able to change server settings and create a role, with at least one table already holding rows, plus a Kafka cluster with Kafka Connect running that you can reach over the network.
Not covered
Taking the initial snapshot of a table too large to read in one pass without long delays or lock contention.
Take a CDC snapshot of a huge table without stalling the pipeline
when the table you're capturing is too large for a single-pass snapshot to run without locking or long delays
Keep a CDC consumer from doubling a row after it restarts
once you're writing a consumer for this topic and need it to survive a restart without processing the same change twice
Make a deleted row vanish downstream with a tombstone
for handling the tombstone so a deleted row actually disappears from a downstream store, not only from the topic
Checked 26 Aug 2026
Part of the Change Data Capture (CDC) cookbook