Serverless: 6 Years Later
summary
Built a production serverless startup on AWS Lambda and now shipping Cloudflare Workers, with the original product still running after acquisition without regular maintenance. Development centers on tiny modules (100 lines of code typical), adding features by adding functions and connections, deploying to prod behind tenant feature flags for safe rollouts, minimal unit tests and no staging, observability and staged internal accounts for validation. Cost model is per-invocation billing so COGS scale at most linearly, giving precise pricing control and much lower ops overhead.
AI-written summary · about
chapters (26)
- 0:00Presentation
- 2:42Serverless
- 3:26What is Service
- 5:30Serverless Execution
- 6:59Functions as a Service
- 12:15Application Development
- 13:34EventDriven Architecture
- 16:56Adding Features
- 19:59Replacing Features
- 21:26Testing
- 24:25Staging
- 25:32Testing Production
- 27:04No More Unit Tests
- 27:52Using Tenants Effectively
- 28:30Local Testing
- 29:18Feature Toggling
- 30:39The Process
- 33:05Programming in the Small
- 33:56Staging Environments
- 36:09My Experience
- 36:57Hygiene
- 38:44Safety Critical
- 41:05Paged
- 41:34Financial Benefits
- 42:09The People
- 46:09Whats next
related talks
29m2 Years, 0 Servers: Lessons learned from running a startup 100% on Serverless
58mServerless - the brief past, the bewildering present, and the beautiful (?) future
47mServerless in production, an experience report
54mServerless Architectures: no servers allowed!
05m#Serverless + #Devops = BFF
39mServerless in production, an experience report
29mGoing Serverless
01h 02mServerless in production, an experience report