Apache Kafka® Broker Failures & Other Outages

This post details the impact of Kafka broker and cluster failures on applications, emphasizing the need for robust operational practices and application-level resilience. It outlines best practices for preventing outages (redundancy, distribution, replication, monitoring, upgrades, automation) and discusses how different application architectures (entry points vs. polling, synchronous vs. asynchronous sends) are affected by extended outages. It also describes the characteristics of extended outages and how client configurations influence error handling, providing a framework for designing applications that can gracefully handle service interruptions.