Currently offset creates as a concatenation of connection.uri, database and collection parameters.
https://github.com/mongodb/mongo-kafka/blob/master/src/main/java/com/mongodb/kafka/connect/source/MongoSourceTask.java#L309-L317

This approach has a couple of disadvantages.

1) such offset could contain mongo username and password, which will be written to kafka connect offsets topic. I guess it's not very secure.
2) if the mongo password or username was changed, the offset will be lost.

Is it make sense to use a static partition name? Or in the future, there is could be more than one partition?