Migrating to Serverless - an experience report
summary
Migrated a large collaboration platform from Heroku to AWS Lambda, cutting operational costs and improving scalability while deleting about one third of application code. Uses Lambda, API Gateway, S3, Cognito and CloudWatch for centralized logging, per-function versioning and A/B testing; billing is per-100ms, exposing fine-grained cost and performance per function. Built small single-purpose exporters (PDF, SVG) and a Web API; works but has limits: ~100–200ms extra latency, 5-minute execution ceiling, no enterprise SLA and potential vendor lock-in.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 6:57Reserved capacity
- 8:15Freezing up
- 12:35Three types of toys
- 13:55Services
- 15:11Deployment
- 16:02Create API
- 18:55AWS Console
- 20:41Scaling
- 28:11Time to recover
- 29:26Multiversioning
- 30:46Version push
- 34:08Embrace the platform
- 34:40Authorization
- 35:18Distributed Authentication
- 40:56Stateless
- 46:09Strengths
- 51:54Weaknesses
- 54:03Other services
- 58:15Opportunities
related talks
39mServerless in production, an experience report
01h 02mServerless in production, an experience report
54mServerless Architectures: no servers allowed!
45mPatterns & practices for developing enterprise-scale serverless applications
42mServerless Tales from the Trenches
47mServerless in production, an experience report
42mAWS Lambda from the trenches
21mMigrating Legacy Backends to Serverless in Parallel
