Currently it commits offset of the last record so there is persistent lag = 1 in partition.

Steps to reproduce:
- Run Kafka connector.
- Produce message to respective topic so it's handled my Kafka connector.
- See the lag is 1 on a topic partition.
- Restart Kafka connector.
Result:
- After connector restarted the message is handled again.
Expected:
- After connector restart the message is not handled again.
┆Issue is synchronized with this Jira Task by Unito