Serverless in production, an experience report
summary
Company moves from EC2 to an event-driven, service-oriented stack centered on AWS Lambda, running 170 Lambda functions in production and paying roughly 5% of comparable EC2 costs. Deployment cadence rises from about 4–6 releases/month to 80–100 deploys/month. Concrete builds include a search pipeline via Kinesis → Lambda → CloudSearch, a BigQuery analytics stream (live in 1.5 days), a Redis-backed timeline using SNS for retries, and SSM for secure config. Monitoring uses CloudWatch logs, ES shipping and X-Ray traces.
AI-written summary · about
chapters (14)
- 0:00Introduction and background
- 1:35Legacy system challenges
- 4:45Transition to serverless architecture
- 9:04Rapid feature delivery with Lambda
- 14:29Choosing a Lambda framework
- 17:20Testing strategies for Lambda
- 25:14Continuous delivery setup
- 29:11Monitoring and logging in Lambda
- 33:11Debugging distributed systems
- 35:17Performance tracking with X-Ray
- 37:04Metrics and alerting pitfalls
- 41:22Optimizing Lambda performance
- 42:45Configuration management
- 47:00Q&A and closing remarks
related talks
39mServerless in production, an experience report
01h 02mServerless in production, an experience report
44mServerless Architecture on Production: Good, Bad and Ugly Parts
42mAWS Lambda from the trenches
01h 00mServerless Architecture - Tales from a world without servers
54mServerless Architectures: no servers allowed!
01h 04mMigrating to Serverless - an experience report
23mFrom 1 to 101 Lambda Functions in Production: Evolving a Serverless Architecture–Slobodan Stojanovic
