Generating Deterministic, Synthetic Workloads for Stream Processing

Do you want to kick the tires on Apache Flink®? Do you want to make sure it can perform well for your target use case? How can you do that?  

Unlike batch systems, repeatably testing stream processors is not as easy as downloading files out of an S3 bucket. You need continuous streams of data that reflect the current wallclock time, have commonality across their identifiers, and have a bit of real-world jitter like delayed arrival.  

In this talk, we look at the design techniques you can use to test these systems yourself. We'll cover how to test with deterministic randomness, deterministic time, stable event ordering, and more.  

Attendees will walk away with the tools to exercise stream processing systems with repeatable synthetic workloads.