Palm & Pine Apparel
Rebuilt a seasonal apparel retailer's checkout infrastructure to survive Black Friday traffic without the site falling over.
Industry
Retail & E-commerce
Timeline
8 weeks
Team
2 engineers, 1 DevOps
Category
Cloud & DevOps
Where things stood
Palm & Pine's site had gone down twice during past Black Friday sales because checkout couldn't handle the traffic spike, costing a full day of peak-season revenue each time.
How we got there
Checkout service isolation
Split checkout into its own scalable service so a traffic spike on the homepage or product pages couldn't take down the ability to actually complete a purchase.
Autoscaling and load testing
Set up autoscaling tied to real traffic thresholds and load-tested against 8x last year's peak before the season started, not after.
Real-time monitoring during peak
Built a live dashboard the team watched during Black Friday weekend itself, with alerting tuned to catch a problem in minutes, not after customers complained.
What changed
0
Checkout outages during the following Black Friday weekend
8x
Peak traffic load-tested for ahead of the season
~99.98%
Checkout uptime during peak weekend
Tech stack
AWS
Hosts the isolated checkout service and autoscaling infrastructure.
Kubernetes
Scales checkout capacity automatically as traffic climbs through peak weekend.
Terraform
Defines the scaled infrastructure as code, tested ahead of the season.
Redis
Caches cart and session data so checkout stays fast under load.
Prometheus/Grafana
Powers the live dashboard the team watched during Black Friday weekend.
Docker
Packages the checkout service consistently across the autoscaled cluster.
