Kafka Summit SF 2019 just ended, and videos and slides are uploaded. Besides the keynotes, the following are the top 7 talks I like and some probably I will go through them again:
1. How Kroger embraced a “schema first” philosophy in building real-time data pipelines
Experience in real world about using schema registry and facilitating data usage between business people, developers and data analysts. The tech stack includes schema-registry, gitlab CI and their own services. (Hope that will be open sourced).
2. Event-Driven Model Serving: Stream Processing vs. RPC with Kafka and TensorFlow
Very useful summary for the pros and cons of the following two alternatives: (1)analytic models are deployed in the application (2)they are hosted in a remote model server
3. Please Upgrade Apache Kafka. Now.
Very useful talk to convince people to prioritize kafka upgrade and the tips to upgrade. Helpful to go through it when upgrade is required, or your CI/script supports upgrade already?
4. Streaming Apps and Poison Pills: handle the unexpected with Kafka Streams
Practical recipes for using KStreams when deserialzation error happened
5. Static Membership: Rebalance Strategy Designed for the Cloud
Helpful to understand rebalance, and how the new feature will improve it
Why stop the world when you can change it? Design and implementation of Incremental Cooperative Rebalancing
https://www.confluent.io/kafka-summit-san-francisco-2019/why-stop-the-world-when-you-can-change-it-design-and-implementation-of-incremental-cooperative-rebalancing
6. Kafka on Kubernetes: Does it really have to be “The Hard Way”?
Interesting and fun demo of deploying kafka ecosystem on k8s using Confluent’s Kafka Operator
7. Kafka 102: Streams and Tables All the Way Down
Very clear and understandable explanation of streams and tables, and why join requires co-partitioning
