Off the top of my head, IIRC each connector will have an entry on the topic with a key based on the connector name plus other data specific to the connector (in the blog example it’s the name of the file being ingested; for JDBC connector I’d expect it to be the table name, etc).
So yes - I believe this method should work even with multiple connector tasks. I’d recommend testing it properly first though. If you bump up the log levels in the Kafka Connect worker you should be able to see enough detail of what’s being read from the topic to check things are behaving.