// the find
zhisheng17/flink-learning
flink learning blog. http://www.54tianzhisheng.cn/ 含 Flink 入门、概念、原理、实战、性能调优、源码解析等内容。涉及 Flink Connector、Metrics、Library、DataStream API、Table API & SQL 等内容的学习案例，还有 Flink 落地应用的大型项目案例（PVUV、日志存储、百亿数据实时去重、监控告警）分享。欢迎大家支持我的专栏《大数据实时计算引擎 Flink 实战与性能优化》
A Chinese-language Flink learning repository pairing ~60 chapters of open-sourced book content with runnable Java examples for connectors (Kafka, ES, MySQL, HBase, Redis, ClickHouse, and more), plus a CDC sub-project. Aimed squarely at developers who want working code to copy-paste while reading the accompanying blog series, not at contributors or people building production libraries.
The connector coverage is genuinely wide — Kafka-to-X examples exist for a dozen sinks, each in its own Maven module so you can build just what you need. The Q&A dump in the README is surprisingly useful: 200+ real production questions about checkpoint failures, watermark edge cases, slot tuning, and YARN HA that you won't find in the official docs. The open-sourced book chapters (books/ directory) are more structured than most blog series and walk from DataStream basics through source-code internals. CDC sub-project covers MySQL, Postgres, Oracle, MongoDB, and OceanBase in one place.
Everything is pinned to Flink 1.14.2, which is years behind; several APIs (especially Table API and the new Source interface) changed enough that copying examples directly into a current project will break. A significant chunk of the README links go to a paywalled 知识星球 platform — you can see the question but not the answer without paying. The repo is Chinese-only throughout, which is fine for its audience but worth knowing if you can't read it. No tests anywhere in the codebase: every example is a run-it-and-see demo, so you have no safety net when adapting code and no signal about which examples are still working.