Fighting Covid with Serverless and JavaScript
summary
NearForm builds and open sources the Covid contact-tracing stack at https://github.com/covidgreen and runs it on AWS using React Native frontends, Node.js backends with Fastify, API Gateway + JWT validation at the edge, ECS/Fargate and Lambda workers. An exposures lambda precomputes files every five minutes which are served from S3 to millions of phones (peaks ~15M downloads/day) and a Secrets Manager hot-spot hits ~4,000 requests/sec. Lambdas scale and cut cost but cause failure modes (missing async, cold-starts); Terraform modules and monitoring prove tricky; vertical regional slices and simple flows work best.
AI-written summary · about
chapters (10)
- 0:00Introduction and background
- 3:54Contact tracing problem and solution
- 7:05Serverless architecture overview
- 10:57When to use and avoid Lambdas
- 15:36Contact tracing workflow and S3 scaling
- 19:47React Native and open-source approach
- 25:04Lambda patterns and SMS handling
- 29:00Lessons learned from the project
- 33:15System design and monitoring tips
- 35:01Q&A and contact information
related talks
37mServerless from Scratch
33mBuilding Open-Source Cloud-Native Apps in JavaScript
51mArchitecting for Hyperscale with AWS: Navigating Real-world Serverless Challenges
39mBuilding Highly Scalable Microservices with Serverless
44mServerless: Giving superpowers to frontend developers
49mTales From the Serverless
29mCreating a cloud native serverless web app – An AWS project reference
29mGoing Serverless