Personalisation at Scale

The client had a big-bang launch event, and the endpoints needed to be ready to face an instant spike in global traffic.

We delivered a multi-region solution that provided low latency to all users, while providing the scale and resilience the event needed.


Global traffic spike requiring large capacity

Need to meet target response times for endpoints

Personalised responses reducing the ability of CDN to offload

Technologies & Techniques

Lambda for Compute, using provisioned concurrency

Automated Multi-region CodePipeline

DynamoDB replication used to place data at all regions

Route53 geo-routing to direct traffic

Cloudwatch Monitoring & Dashboards


Endpoints satisfied the loading at target latency