Case Study: Fetch achieves scalable live streaming for 3.2 million viewers with Amazon Web Services

Fetch Cuts ML Processing Latency by 50% Using Amazon SageMaker

Fetch, which already used Amazon Web Services as its primary cloud provider, needed to support a major live-streaming event and handle traffic spikes far beyond its normal app load—potentially 100 times the usual requests per second. To prepare for millions of viewers, Fetch worked with AWS services including AWS Enterprise Support, AWS Countdown, Amazon CloudFront, Amazon S3, Amazon MSK, AWS WAF, and AWS Shield Advanced.

Amazon Web Services helped Fetch validate its architecture, plan capacity, and test scaling ahead of the event through daily stress tests and realistic simulations. Using Amazon CloudFront Origin Shield to reduce origin load and Amazon MSK for near real-time winner selection, Fetch successfully delivered the stream to 3.2 million viewers, improved security, and avoided backend overload; the testing also uncovered and resolved issues before launch, helping the event run smoothly.

Fetch

Alec Stashevsky

Applied Scientist