New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
Why did the whole partition stop moving over one bad row?
What you'll have at the end
A dead-letter topic wired into the consumer: the record that can't be processed is diverted there with enough context to debug it, while every other record in the partition keeps moving.
You need
A Kafka consumer that already connects to a topic, deserializes each record, and does real work with it, such as writing to a database or updating a running total.
Not covered
Kafka Streams' own built-in dead-letter routing or a Kafka Connect sink connector's built-in dead-letter setting, since both wire failures differently than a plain consumer does.
Leans on
Commit Kafka consumer offsets so a crash doesn't skip or repeat messages
covers committing offsets safely through an ordinary retry or crash, for the case where every record can eventually be processed and the only real question is surviving a restart.
Every consumer breaks without a Kafka Schema Registry check
if the records landing in your dead-letter topic are mostly schema-shaped, a field type or a required field changed upstream, catching that before it reaches the consumer is the better fix.
Checked 25 Aug 2026
Part of the Apache Kafka cookbook