Dead Letter Topic #01
Story Horse?
Welcome to the first edition of Dead Letter Topic, a monthly series here on Topic:Igor where I’ll share interesting tools, libraries, frameworks, and resources from the broader event streaming and real-time data ecosystem. These are the messages that didn’t quite fit into the main Kafka stream but still deserve to be delivered: projects like Apache Spark, Apache Flink, Pulsar, and many others that complement or compete in the same space. Expect one carefully curated post per month, no spam, just useful links and quick thoughts on what’s worth exploring next in this ever-growing world of data in motion.
Flink
Building a Flink Self-Serve Platform on Kubernetes at Scale
Real-time Twitch chat sentiment analysis with Apache Flink
Bridging Flink SQL and Custom Java Pipelines with the Decodable SDK
Building Apache Flink Applications in Java
Parallelizing High-Lag Backfills with Flink.
Fluss
Accelerating Stream Processing in Production
What the Fuss with Fluss: Flink 2.2 Delta Force
Evolving To Streamhouse: Apache Fluss x Iceberg
A fraud detection pipeline with Streamhouse
WALs, Changelogs, and Real-Time State
Spark
A 9-minute simple explanation of Spark Shuffle
Real-Time Mode: Ultra-low latency streaming on Spark APIs without a second engine
Debezium
Introducing reusable Connections in the Debezium Platform
Extra
Seamless streaming to Amazon S3 Tables with StreamNative Ursa Engine
Dynamic consistency boundaries
Why AI Agents Need a Protocol-Flexible Event Bus
Streaming Protobuf to Managed Iceberg on Databricks: A Practical Guide
Let me know if you discover anyone interesting.
Any other link or content that I missed? I’d love to hear them.