Peak Performance ⛰️ Defining an Architecture for Managing Traffic during BFCM

With Black Friday/Cyber Monday sales just around the corner, eCommerce sites around the globe will see site traffic spike far beyond their daily average, pushing their websites to the limit. According to a recent report by Adobe , Cyber Monday alone in 2024 is set to generate over $13 billion in the US. During any peak sales period, it’s crucial to have a robust site that can effectively handle hundreds of thousands (and sometimes millions) of user requests and transactions. If the customer experience is disrupted by downtime or slow load times, this can result in missed opportunities, lost sales and sometimes even a decline in customer loyalty. Here’s how our engineering teams at Rotate° address some of these challenges.

Building for scale: leveraging edge networks

Each platform within your tech stack has a usage limit, and downtime or errors occur when those limits are reached and the servers can’t handle additional requests. Monolithic platforms that rely on centralised servers usually have more limitations and we tend to see a higher amount of outages for brands with a monolithic architecture compared to those who have taken a headless or composable approach.

Unlike central servers that handle all website and app requests from a single location, an edge network is distributed across multiple global locations or Points of Presence (PoPs), allowing requests to be distributed geographically based on user location and reducing the distance that data has to travel. These networks can store data from images and videos to pages and JavaScript files at any of those PoPs, making it an effective way to handle high traffic volumes whilst maintaining performance.

Managing data flow at scale

Having an event-driven architecture to manage the scale during peak trading times is essential to ensure smooth operations and customer satisfaction. During traffic surges, third-party integrations including payment gateways, subscription platforms and inventory services may struggle under heavy load, causing errors or delays. A server-management tool can help process data effectively and ease the pressure on third-party providers. At Rotate°, we like to use Hookdeck for 3 key reasons:

1. Scalability – it is purpose-built to handle the significant surges in webhook traffic that BFCM brings, automatically scaling to ensure every purchase order, inventory update and shipping notification is successfully captured and processed. 

2. Precision – its advanced traffic management features allow the pace of event delivery to be regulated through rate limiting. By minimising the risk of overwhelming third-party systems and managed infrastructure, brands can maintain optimal performance and reliability even during peak demand. 

3. Reliability – it has robust backup and event recovery capabilities that allow it to store events to be processed later if a specific system goes down temporarily. This acts as a safety net and ensures no transactional data is lost.

Optimising data management

During peak periods, websites experience a constant demand for product data such as stock levels, pricing and product details. Managing this directly through your eCommerce platform's admin API puts strain on the system as more requests flood in. Taking a composable approach helps reduce this pressure. By using a headless CMS almost like your Product Information Management (PIM) system, your eCommerce platform can focus on other critical tasks such as order processing, payment and checkout. As a result, the website stays responsive and product data is maintained and kept up-to-date without overloading the backend API.

Ready for peak?

Last Cyber Monday, Cloudflare reported that at 4:10pm it processed a peak of 80 million HTTP requests per second. As Black Friday and Cyber Monday approach and many sales are already beginning, technical teams should assess what strategies they have in place to support them during high traffic periods. Leveraging edge networks, managing data flow and optimising product data effectively all help to build a robust system capable of delivering a best-in-class shopping experience for customers all year round.