We tested kafka connect and felt it was hard to work with. Had to setup 3 steps a) stream changes to kafka b) sync raw changes from kafka to the target data-store c) normalize the raw changes to the target. Overall it wasn’t that trivial and there were multiple moving parts. PeerDB should be much simpler than that - create peers and create a mirror using simple SQL commands. This takes care all the above steps.

Also my understanding is the kafka connect means you need to use debezium. Sharing a thread which captures a few benefits of PeerDB vs Debezium - https://news.ycombinator.com/item?id=36898413