Christoph Schubert
Christoph is a Solutions Architect with Confluent. Before joining Confluent, he developed streaming applications for the financial services industry and mobile applications for about everyone else. He holds a PhD in Mathematics and has a long-lasting interest in developing performant and secure software systems.
Session
Kafka Streams is a library for developing streaming application with Apache Kafka.
We will discuss best practices for developing a production-ready Kafka Streams application and for running it smoothly in production.
After reviewing the fundamentals of stateless and especially stateful programming with Kafka Streams, we will address the following questions:
-
How to prepare your application for seamless failover?
-
How to deal with the ever-growing table anti-pattern and properly implement TTL?
-
How to prevent resource-leaks when dealing with RocksDB-based state stores?
-
Which metrics to monitor?
-
How to size your runtime environment?
-
What should we keep in mind when deploying Kafka Streams on Kubernetes?
-
How to best deal with evolving data models?