Intermediate8 min
Triage a Dead Letter Queue Flood
Practise sampling and classifying a DLQ flood instead of reacting to it record by record.
This lab is a paper-and-CLI exercise, not a simulator one. A Kafka Connect sink connector has errors.tolerance=all and a dead letter queue (errors.deadletterqueue.topic.name). It receives records that failed in the converter or a transform, and its topic has grown from near zero to 25,000 records in an hour. The better move is to sample first and let the pattern show what broke.
The steps use the simulator and name the control to use. The outcomes they describe are what the simulator models; the sources below are what Apache documents.
Steps
0 of 5 steps completed
Verification
You can explain in a minute what the single cause was, how you found it without reading 25,000 records, and what you would check before replaying.
The interview angle
The answer to give is: sample, classify, read one representative record, fix at the source. The dead letter queue holds records that failed in the converter or in a transform, for sink connectors only.
Sources
- Kafka Connect user guide (4.2 documentation) (opens in a new tab)
- Kafka Connect configs (4.2 documentation) (opens in a new tab)
- KIP-298: Error Handling in Connect (opens in a new tab)
Checked against Kafka 4.2 documentation on 11 October 2026. Draft: not yet reviewed by a Kafka specialist, and releases move on, so confirm against the current documentation before you act on it.