Migrating Legacy Backends to Serverless in Parallel
summary
Migrates a Ruby 1.8 Rails backend to functions by extracting endpoints into Ruby 1.8 → AWS Lambda handlers behind API Gateway, deployed with the Serverless Framework, while leaving heavier parts on ECS and RDS; traffic routes split so new Lambdas and existing services run in parallel. Uses MongoDB for schema‑less extracts, VPC to reach RDS, and Step Functions for orchestration. Shows a cost example (1,600 requests/day, 200ms: EC2 ~$3/day vs Lambda ~$0.05/day) and documents real pitfalls: cold start vs warm start, library bloat, duplication and vendor lock‑in.
AI-written summary · about
chapters (10)
- 0:00Start
- 2:22WHY MANAGING SERVERS IS SICK
- 5:03SOME INTERESTING STATS
- 6:09DIFFERENT VENDORS
- 6:38COMPARISON WITH EC2
- 7:20TOOLBOX (AWS)
- 11:25SWISS KNIFE(S)
- 14:20HOW WE STARTED MIGRATING
- 16:24SOME COOL STUFF
- 17:17IS IT ALL GOOD???
related talks
01h 04mMigrating to Serverless - an experience report
39mServerless in production, an experience report
45mPatterns & practices for developing enterprise-scale serverless applications
42mAWS Lambda from the trenches
43mServerless in Java Lessons learnt
01h 02mFrom Monolith to Serverless
42mServerless Tales from the Trenches
31mA quick intro to AWS Serverless