TechPrufer
Get started
Shopify
Shopify logoShopify · Feature trail

Unified Edge Traffic Management

2 posts · Feb 2017 – Dec 2025

What this trail covers

Shopify has unified its edge network infrastructure with the introduction of Möbius and Shopify Front End (SFE). Möbius handles TLS and TCP termination close to the client, providing improved performance and DDoS protection. SFE acts as a unified routing layer behind Möbius, dispatching traffic to various applications and services. This initiative standardizes traffic paths, simplifies observability, and automates edge configuration for application owners, enabling faster feature deployment and improved resiliency. This post details the initial implementation of a scriptable load balancer using Nginx and Lua to handle high-write traffic during flash sales, introducing a leaky bucket algorithm and a queue page to manage customer experience and platform stability.

All Shopify trails
  1. Behind the build: 2025's BFCM globe in a pinball machine 🌎🕹️ - ShopifyThis post details the technical implementation of the 2025 BFCM Live Globe, which was built as a pinball machine. Key technical aspects include: a three.js app with react-three-fiber, server-sent events for real-time order streaming, custom shaders for dot matrix displays, 2D physics simulation using rapier2D for performance, VR support via react-three/xr, and the display of real-time Shopify infrastructure stats on an in-game monitor. It also touches on design decisions like prioritizing 2D physics over 3D for performance and the modular design of interactive minigames within the pinball machine.
  2. Möbius: Shopify’s Unified Edge - ShopifyThis post details the development and implementation of Möbius and Shopify Front End (SFE) to unify Shopify's edge network. Möbius handles TLS and TCP termination at the edge, providing improved performance and DDoS protection. SFE acts as a unified routing layer behind Möbius, dispatching traffic to applications and services. The initiative standardizes traffic paths, simplifies observability, and automates edge configuration for application owners.
  3. How Shopify Dynamically Routes Storefront Traffic - ShopifyThis post details the implementation of dynamic routing rules using OpenResty and Lua to manage traffic shifting between legacy and new storefronts. It describes the evolution of rule management from Lua files to a JSON payload in a control plane, manipulated via a chatbot (spy). Key configuration options for rules include rule name, filters, shop IDs, render rate, verify rate, reverse verify rate, and self-verify rate. The technical implementation highlights the use of nginx directives (rewrite, header filter, log phase) and Lua modules for rule matching and asynchronous verification requests. The post also discusses the benefits of dynamic rules for removing CI/deployment time and enabling rapid iteration and testing of new features and resiliency scenarios.
  4. Iterating Towards a More Scalable Ingress - ShopifyShopify contributed dynamic configuration to ingress-nginx, enabling it to update application endpoints in-memory via an HTTP endpoint served by lua-nginx-module. This eliminates the need for NGINX reloads during deployments, significantly reducing request latency and improving performance, especially during high-traffic events and deployments.
  5. Surviving Flashes of High-Write Traffic Using Scriptable Load Balancers (Part II) - ShopifyThis post details the implementation of stateless fair queueing to improve fairness during high-traffic events. It introduces the use of timestamps to prioritize users based on their arrival time and a PID controller (specifically, the proportional term) to dynamically calculate a threshold for requests. This threshold is managed independently on each load balancer to maintain statelessness and ensures that requests are processed fairly without relying on a central data store. The effectiveness is demonstrated through graphs showing reduced variance in customer queue times.
  7. Surviving Flashes of High-Write Traffic Using Scriptable Load Balancers (Part I) - ShopifyThis post details the initial implementation of a scriptable load balancer using Nginx and Lua (OpenResty) to handle high-write traffic during flash sales. It introduces the concept of back pressure and a leaky bucket algorithm to manage traffic spikes, preventing platform crashes. A queue page was developed to maintain customer experience, and a mechanism for customers to skip the throttle after initial entry was implemented. The post highlights the limitations of this initial solution, specifically the random reselection in the queue, which is addressed in a subsequent post.