The author, an infrastructure engineer at Contentful, recounts the story of how they implemented multi-region active-active file delivery on Amazon S3 using CloudFront and Lambda@Edge. The goal was to increase availability by serving files from multiple regions, but this introduced new risks, such as the possibility of causing an outage themselves. To mitigate these risks, the team created a production readiness criteria list, which outlined requirements for monitoring, deployment, scaling, and other aspects of the system. They also used TypeScript to write more reliable code and deployed the platform using automated tools. By addressing their engineering gut feelings and operational experience, they were able to build a reliable software platform that increased availability while minimizing risks. The implementation demonstrates the challenges and trade-offs of edge computing and the importance of careful planning and risk management in cloud infrastructure deployments.