At a glance
The challenge
During festive-season and flash sales, traffic jumped many times over normal levels. The store slowed to a crawl or went down entirely at exactly the moment marketing spend was highest.
To cope, the team kept servers oversized all year, paying for capacity that sat idle most of the time.
Our approach
Find the bottlenecks
Load tests showed which parts of the store failed first: checkout, search and the product database.
Re-architect for scale
We containerised the application, added caching and moved static content to a CDN.
Scale automatically
Auto-scaling rules add capacity as traffic rises and remove it when the rush ends.
Rehearse sale day
We load-tested above expected peaks before every major sale.
The solution
The store now scales up automatically when a sale starts and back down afterwards. Pages and images are served from locations close to shoppers, keeping the site fast across regions.
The business pays for peak capacity only during peaks, and the team goes into each sale knowing the platform has already handled more.
Results
- Sale traffic handled without slowdowns or outages
- Lower hosting costs outside peak periods
- Faster page loads through global content delivery
- Load tests before every major campaign
Technology used
- AWS
- Docker
- Kubernetes
- Redis
- CloudFront CDN
- k6 load testing
Facing a similar challenge?
Tell us what you are working on. We will walk you through how we would approach it, with no obligation.