Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about Apache Flink's changelog modes in this 27-minute conference talk that explains how small SQL statement changes can dramatically impact state size, latency, network usage, and compute requirements. Discover the four changelog modes (insert, update before, update after, and delete) and understand how different operators consume and produce these modes throughout the data pipeline topology from source to sink. Explore how primary keys and upsert keys influence changelog modes, and gain practical insights into optimizing data pipelines by managing these modes effectively. Master the concepts that help prevent unexpected performance issues and enhance streaming application efficiency, with real-world examples demonstrating how upsert modes can reduce network traffic and how understanding intermediate operators like changelog normalize can explain state size growth.
