Hi folks. I am currently trying to setup CDC pipeline, from MongoDB (using Debezium) to Flink to JDBC sink table in Postgres. I am able to see inserts and updates being processed successfully in the downstream Postgres table, however delete handling is not working correctly. I have tried changing the 'drop.tombstones' property to 'false' and changing the 'delete.handling.mode' to 'rewrite', 'none', and 'drop', none of which seems to fix the problem. I have also tried using both 'kafka' and 'upsert-kafka' for the source kafka table definition in Flink. I have read some blog posts on this topic but unfortunately most of the examples I've seen with debezium are using a source db like MySQL instead of a document database such as MongoDB.

@Gunnar Morling Any thoughts here?

I found the

Debezium Unwrap MongoDB SMT Demo

in the debezium-examples GitHub repo. This is more or less what I am trying to achieve except with another Flink kafka table as an intermediary. On my original debezium kafka connector, I added the

unwrap

transformation so that the messages generated in the kafka topic can be queried in Flink without being deeply nested. Not sure if that is causing the problem?