Leaving the Data Center: Shopify's Move to Google Cloud and Kubernetes
Why Shopify migrated its core commerce platform out of self-managed data centers onto Google Cloud and Kubernetes, and how it did so without downtime.
Engineering at scale
11 articles · 42 min of reading
Why Shopify migrated its core commerce platform out of self-managed data centers onto Google Cloud and Kubernetes, and how it did so without downtime.
How Shopify built the Liquid templating language to safely run code written by outside theme developers on shared production servers.
How Shopify built Toxiproxy to deliberately simulate latency, timeouts, and bandwidth limits so engineers could test resilience before production did.
How Shopify stayed on a modular Rails monolith and isolated flash-sale blast radius with pod-based sharding instead of splitting into microservices.
Why Shopify moved its Admin and Storefront APIs from REST to GraphQL as the primary way tens of thousands of apps talk to the platform.
How Shopify keeps a single large Rails monolith deployable dozens of times daily using merge queues, canaries, and fast automated rollback.
How Shopify runs enormous volumes of background jobs safely, treating idempotency and graceful interruption as first-class design constraints.
Shopify rebuilt storefront rendering as a dedicated service to cut page load times, separating buyer-facing traffic from the core Rails monolith.
Shopify runs months of scripted load tests and failure drills against production infrastructure to prepare for Black Friday Cyber Monday.
Shopify funded a just-in-time compiler for Ruby instead of rewriting its monolith, turning YJIT into a core piece of commerce infrastructure.
How Shopify layered gradual, static type checking onto a massive dynamically-typed Rails codebase using Sorbet without freezing feature work.