Company
Date Published
Author
Terri Haber
Word count
2103
Language
English
Hacker News points
None

Summary

The text discusses the concept of "adversarial game days" in software development and reliability engineering, which involves intentionally introducing faults into a system to test its resilience and recovery capabilities. New Relic, a software company, uses this approach as part of their site reliability best practices. The goal is to identify potential weaknesses in the team's collective knowledge and improve the mental model of the system. Adversarial game days can be run in various environments, including non-production versions of the system, and tools such as Toxiproxy can be used to inject faults into the system. The exercise should be conducted like an actual incident, with a scribe taking notes and customer-support representatives participating, and a retrospective held after the exercise to discuss what could have been done better and what was learned.