How Shopify handles 16,000 Request per second

Last month Refreshers :

1. Best Practices for Developing Microservices

2. Software Architecture Pattern - Layered Architecture

3. How Enterprise Applications Exchange Data Using Messaging

4. Microservices Design Pattern - Event Sourcing Pattern

5. Improve API Performance 🚀

6. Distributed System Learning Roadmap

---

Problem

Shopify has grown to host around 275,000 merchants and handles approximately 16,000 requests per second on a typical day.

Shopify faced significant scaling challenges due to unpredictable traffic spikes from flash sales, leading to database bottlenecks and affecting overall platform stability.

Evolution of Shopify's Architecture

Initially, Shopify operated as a single-tenant application but quickly transitioned to a multi-tenant model to accommodate growth.

For several years, the architecture saw minimal changes, leading to challenges such as the Flash Sale traffic pattern, which caused significant scaling issues, particularly with the database.

In 2013-2014, Shopify focused on database-level isolation, which significantly improved scalability.

Flash Sale Problem

• Flash sales occur when merchants announce limited inventory sales, leading to unpredictable spikes in traffic that are difficult to manage.

• These events typically result in heavy write traffic to the database, complicating caching and resource provisioning.

• Shopify's architecture needed to be over-provisioned to handle these unpredictable traffic surges.

Shopify Architecture :