Your DLQ has 25,000 unread messages that built up over the last week. Walk through how you'd triage it.

1) Sample a small, recent batch of DLQ records with full headers rather than reading everything. 2) Group and count by exception class to find the dominant failure mode(s) — this usually reveals 1-2 root causes, not thousands. 3) Inspect one representative message per exception class in detail to pin down the exact cause. 4) Fix at the source: correct the producer schema, add a transform, or repair the sink config/credentials — don't just raise retry limits. 5) Before replaying: verify the fix on a sample, check for side effects like duplicate downstream calls, cap the retry count, and keep a way to abort the replay if it goes wrong.

Expected key concepts: DLQ triage, sample first, group by exception class, batch amplification, fix at source, safe replay, kill switch, Kafka Connect