Event-based, serverless app to broadcast real-time traffic alerts
summary
An event‑sourced, serverless system broadcasts transit alerts to Twitter, Facebook, SMS, email, Google Maps and third‑party apps by storing write‑once events in DynamoDB and routing them with EventBridge to Lambda projectors that build query projections. It uses API Gateway and Cognito for ingestion, WebSockets for optimistic UI with live projection updates, and dead‑letter/retry patterns for failed deliveries. Running in production for multiple RATPDev networks, it exposes tradeoffs: DynamoDB event store, EventBridge routing, Lambda projectors — projection lag and debugging distributed workflows are the main pain points.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:26Transport disruption challenges
- 3:33Why serverless architecture
- 5:21Event sourcing decision
- 6:24Event sourcing theory
- 11:03Event sourcing benefits
- 13:10CQRS pattern explained
- 14:13Serverless service design
- 17:26NoSQL database choice
- 19:11Lambda functions in action
- 21:50Projection conditions
- 23:37Event store attributes
- 25:22Projection types
- 26:25Event bridge mechanism
- 27:51Lambda retry features
- 28:55Production deployment
- 29:40System complexity growth
- 31:49Eventual consistency issues
- 33:41Business validation pitfalls
- 36:08Summary and resources
related talks
29mHow to build a social network on serverless
45mBuilding Event-driven Serverless Applications
36mLeveraging Serverless architecture with Confluent Kafka
43mRealtime Serverless Communication with
47mPatterns for Building Event-driven Web & Mobile App Backends
47mEvent-driven Serverless Architecture with DynamoDB
25mEvent driven design and serverless
44mHow to build a social network entirely on serverless