Chaos engineering is a relatively new discipline within software development, mostly aimed at (but not limited to) distributed systems. While often misinterpreted as ‘breaking stuff in production’ (which sounds cool, but doesn’t add much value), I will show that this discipline should be understood differently. It enables developers to make the inherent complexities (‘chaos’) in modern applications insightful, and lets us detect vulnerabilities and unexpected dependencies in the system in a controlled way.

I’ll show how a chaos engineering tool like Gremlin can help us to conduct experiments (e.g. by tweaking latency parameters or time-travelling to Daylight Saving Time). Besides detecting potential problems in our systems (and diminishing the blast radius of those failures), these experiments also provide a perfect opportunity to test and improve our alerts.

DevOps / Software engineers interested in improving system resilience and pro-active spotting bugs and unexpected dependencies.