Choosing Kafka Producer Acks: Balancing Latency and Durability
Learn how to pick the right Kafka producer acks level and configure min.insync.replicas so you can trade write latency against guaranteed durability.
Loading…
ReadMeFeed / Community knowledge
#apache-kafka
Real questions. Useful conversations. Find the people who know your stack.
apachekafkaAchieving Exactly-Once Delivery in Apache Kafka with Idempotent Producers and Transactions
Enable Kafka exactly-once by turning on idempotent producers and using the transactional API. This guide shows the mechanism, a complete Java example, limits, and verification commands.
Cutting Kafka Rebalance Disruption with CooperativeStickyAssignor
Rolling restarts don't have to pause every consumer. Here is how CooperativeStickyAssignor reduces partition reassignment disruption and what to verify before relying on it.
Kafka Log Compaction: How to Enable and Verify State‑Store Topics
Use Kafka’s log compaction to keep only the latest value per key in a topic. This guide walks through enabling the feature, testing it, and troubleshooting common pitfalls.
Why LinkedIn Built Kafka as a Commit Log Instead of a Message Queue
LinkedIn built Kafka as a distributed commit log rather than a message queue. Here's why that one decision enables replay, independent consumers, and massive throughput — and what it costs.
Stopping Duplicate Messages with Kafka Idempotent Producers
Learn how to eliminate duplicate messages in Apache Kafka using the Idempotent Producer. Explore the PID/Sequence mechanism, required configurations, and how to verify success using JMX metrics.
Guaranteeing One‑Time Delivery in Kafka: How Exactly‑Once Semantics Work in Practice
Duplicate messages can break event‑driven systems. This post explains Kafka’s Exactly‑Once Semantics, shows how to enable it with a producer and consumer, and discusses the trade‑offs you’ll face when turning EOS into production.
Kafka's Idempotent Producer: Exactly-Once Per Partition Without the Transaction Overhead
Kafka's idempotent producer kills retry-induced duplicates with one flag — if you respect the config contract and know it only covers a single partition per session.
Configuring Kafka Log Compaction for State Recovery in Stream Processing
Enable and tune Kafka log compaction so stream-processing state rebuilds stay fast: topic configuration, dirty-ratio tuning, tombstone deletes, and verification with console producers and consumers.