TALK ABSTRACT

Tech debt eats at a long running distributed service. This causes fear of augmenting the system with new features, and inevitably leads to heavy ops. In this talk, I present a migration journey to a serverless system, through A/B methodology, and the learnings incurred along the way.

A long running distributed service will soon reach its limits, either with scale or with functionality and it is time to upgrade the house. This talk addresses the challenges of migrating a long-running distributed service to a serverless system. It outlines a low-risk A/B experiment methodology for migration, emphasizing traffic measurement, system evaluation, decision-making, optimization, and iteration to ensure real-time results and successful infrastructure transition, all when the user is none the wiser. And as an SRE, is it blissful while you handoff your infrastructure to a serverless system, or is more control a better option?