On Black Friday 2025, Shopify's Storefront served billions of requests on Falcon—23 million RPM at peak, zero dropped requests. One year earlier, we ran three different web servers and had no idea which parts of our 15-year-old Rails codebase were thread-safe.

This talk brings together three perspectives on migrating a massive Rails application to async Ruby in production:

Samuel Williams discusses production reliability: worker health monitoring, proactive memory management, and fiber profiling to detect code blocking the event loop.

Marc-André Cournoyer dives into "long tasks": an approach that lets CPU-bound Rails apps handle concurrent I/O without performance degradation. GraphQL @defer went from blocking workers for 30+ seconds to nearly free.

Josh Teeter shares the infrastructure wins: the Outbox system (hundreds of Kafka connections → one per pod), debugging issues that only surfaced at BFCM scale, and orchestrating a zero-downtime rollout.