davsclaus opened a new issue, #3360: URL: https://github.com/apache/camel-kamelets/issues/3360
The catalog has no change data capture (CDC) source: no Kamelet sends the changes of a database table as they happen. It has 9 SQL Kamelets, which all poll with a query. A `debezium-postgresql-source` Kamelet, on the Camel `debezium-postgres` component, would fill that gap. MySQL and SQL Server could follow the same pattern later. A Kamelet is most useful when the changes go somewhere other than Kafka (HTTP, S3, a search index, a queue), and in Pipes and Kaoto; for Kafka itself there are Debezium Server and Kafka Connect. - **Output:** one message per row change, with the row after the change (before it, for a delete) as JSON. The operation (create, update, delete, read for the snapshot), the table and the key are headers. - **Offsets:** Debezium keeps its position in a local file by default, which is lost in a pod without a volume. The Kamelet exposes the file, and the Kafka offset store as an option, so that a restart goes on where it stopped. - **Options:** the component has about 146 options. The Kamelet takes a curated set (connection, credentials, the tables to include, the snapshot mode, the slot and publication names) and says how to reach the rest. - **Security:** the password is marked as a credential, and the replication slot and publication are named by the operator, not by message data. - **Citrus tests:** a PostgreSQL container with `wal_level=logical`. An insert, an update and a delete each arrive with the right operation and body, and a restart does not send them again. _Claude Code on behalf of Claus Ibsen_ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
