Kafka Streams in Action (William P. Bejeck Jr.) (Z-Library)

# Kafka Streams in Action (Second Edition) — Reading Guide ## 【One-Line Pitch】 A practical, hands-on guide to building event-driven applications and microservices with Kafka, covering everything from brokers and clients to stream processing—ideal for Java developers who want to move from theory to working code quickly. ## 【Book Arc】 - **Opening (~0%–10%)**: Introduces the Kafka event streaming platform, contrasting messages (tactical, machine-to-machine) with events (business-critical data worth preserving), and sets up the Docker-based development environment used throughout. - **Early (~10%–23%)**: Walks through broker fundamentals—produce/consume requests, partitions, replication, and metadata management—then pivots to schema design with Avro, Protobuf, and JSON Schema, including Schema Registry integration and subject name strategies. - **Early–Middle (~23%–39%)**: Covers schema references for handling multiple event types per topic, then dives deep into Kafka clients: producer configurations, custom partitioners, consumer groups, offset committing, and the two-level heartbeat/session timeout system. - **Middle (~39%–48%)**: Explains consumer group mechanics in detail—group coordinators, eager vs. cooperative rebalancing, and the stop-the-world cost of rebalances—then introduces exactly-once delivery semantics with idempotent and transactional producers. - **Late (~48%–end)**: Moves into Kafka Connect for integrating external systems, including Single Message Transforms, sink connectors, and building custom connectors—the bridge between Kafka and the rest of your architecture. ## 【Key Takeaways】 - **Events are business-critical data, not transient messages** (Early): Kafka's value proposition is preserving events for replay and audit, not just passing messages between services—this mindset shapes all downstream design decisions. - **Schema management is non-negotiable** (Early): Avro, Protobuf, and JSON Schema each have trade-offs; Avro uses JSON definitions with defaults, Protobuf relies on field numbers and type-based defaults, and JSON Schema is most readable—pick based on your ecosystem. - **Subject name strategies control schema evolution** (Early): TopicNameStrategy, RecordNameStrategy, and TopicRecordNameStrategy each balance flexibility for multiple types per topic against version control—choose deliberately based on whether you need multiple event types or cross-topic versioning. - **Schema references constrain multi-type topics** (Early): Using Avro unions with schema references lets you have multiple event types per topic while limiting them to a defined set—critical for maintaining order and sequence of related events. - **Custom partitioners give fine-grained control** (Middle): When default partitioning doesn't fit your data patterns (e.g., routing special orders to a single partition), implementing a custom partitioner is straightforward and powerful. - **Consumer group coordination has real costs** (Middle): Rebalances are "stop-the-world" operations; cooperative rebalancing (Kafka 2.4+) reduces downtime, but you must understand group coordinator mechanics to design resilient consumers. - **Exactly-once delivery requires transactional coordination** (Middle): The producer epoch mechanism prevents zombie producers from corrupting transactions—if a producer is fenced, you must close and restart it, making failure handling a design consideration. ## 【Reading Tips】 - **Skim the broker internals** (~10%–19%): You need the mental model, not the implementation details—focus on produce/consume request flow and replication basics, then move on. - **Deep-read the schema chapters** (~19%–39%): Schema design and Schema Registry configuration are where real-world Kafka projects succeed or fail; pay special attention to the subject name strategy comparison table and schema reference examples. - **Study the consumer group section carefully** (~39%–48%): The group coordinator, rebalance protocols, and offset committing are the most conceptually dense material—work through the figures and examples slowly. - **Use the code samples as your primary learning tool**: The book references specific classes (e.g., `BaseProducer`, `TransactionalProducerConsumerTest`) in the accompanying source—clone the repo and run the tests to solidify understanding. - **Skip the acknowledgments and foreword** (~0%–3%): They add context but no technical content; jump straight to Chapter 1. ## 【Coverage Limits】 This guide covers the first half of the book (through Kafka Connect introduction). The excerpts do not cover Kafka Streams DSL specifics, state stores, KTables, or the stream processing portions promised by the title—those appear in later chapters not included in this sample. ##