Scale, Brownouts, and Resiliency Challenges

What is a Principal Engineer at Amazon? With Steve Huynh

Topic 3 of 11

Published Jul 9, 2025 · Clip 8:58–16:40 (7:42)

Steve describes Amazon-scale engineering challenges: the Prime Video gateway page fans out a single request into hundreds of downstream calls, producing tens or hundreds of thousands of requests per second for internal microservices. He introduces 'brownouts'—outages where services are reachable but time out or return partial or bad results—and the difficulty of dependency chains and recovery without re-overwhelming a recovering service. He notes that transactional semantics and live sports raise the stakes beyond best-effort availability, and explains why startups love hiring from Amazon: developers are handed a pager and must operate the software they write.