I had a question about the recommended approach to deal with fatal business errors during Kafka streaming.
Scenario: