In today’s world, we deal with essential complexity in our systems. As we get closer to a microservices like architecture, we could have multiple layers of dependencies, where each service on its own might have been functionally and abilities tested, but harder to simulate when one service or dependency might go rogue in a specific circumstance in an entire workflow. This interaction (or just behavior change in different environments, configurations) is too complex for humans to predict.
The discipline of Chaos Engineering is to create these ‘random’ scenarios (eventually in an automated fashion) and build the resiliency into the system as a whole, while increasing the velocity at which value is delivered to consumers. In this talk, we will discuss the principles for Chaos Engineering, how to plan for introducing chaos in your organizations and why a culture of DevOps is essential for it to succeed. We will also talk about where chaos experiments should be run, using business metrics to measure an experiment’s results, permeating the results to all teams as well as how to scale it across a larger organization by enabling each team to run their own experiments.
Finally, we will also discuss our team’s journey to Chaos Engineering at Cerner, including experiments we have run and what we learnt from them. The goal will be for attendees to walk out excited about Chaos Engineering and also with ideas on starting on the path to adopt it in their organizations.