Kafka: acks=all is not durable without min.insync.replicas

acks=all waits for every in-sync replica to acknowledge. If replicas have fallen out of the ISR, 'every in-sync replica' can mean just the leader - and a single broker failure loses acknowledged writes.

• Kafka: linger.ms=0 does not mean every record is sent alone
• Kafka: an eager rebalance stops the entire consumer group
• Kafka: log compaction never removes the last record for a key, including a tombstone
• Kafka: consumer lag on one partition is skew, not a scaling problem
• Kafka: a read_committed consumer stalls behind one open transaction
• Kafka: a non-idempotent producer reorders on retry

---

Run this yourself in the Online Java Compiler, spin up a live REST API in the API Sandbox, or practise with Java interview questions.

Published 2026-08-25

© Java Coding Hub · About · Contact · Privacy · Terms