A Kafkaesque Series of Events

Explore a deep-dive investigation into a complex concurrency bug in Rust that evaded the language's type system. Delve into the asynchronous messaging systems at OneSignal, processing billions of events daily, and uncover how a series of unfortunate coincidences led to a near-total work stoppage. Learn about Kafka, event streaming, and data-intensive applications while gaining insights into observability, subpartition processing, and problem-solving in large-scale systems. Discover the challenges of concurrent programming and the importance of thorough debugging in this informative conference talk from YOW! 2022, presented by Lily Mara, Engineering Manager at OneSignal.

Intro
Once upon a time
Kafka
Issues
Subpartition processing
Concession
Review
Goals
Observability
OneSignal
Why is that a problem?
Q&A
Outro