What We Learned Going Serverless
summary
Serverless implementation uses AWS services (API Gateway, AWS Lambda, S3, DynamoDB, Cognito, SQS, SNS, CloudFormation) to ingest 35,000 daily files; a load test of 1,000 small files triggers 9,000 function executions and never exceeds 68 queued messages, while a 3,000 large-file test (each 91 GB) averages 15 seconds per file. Infrastructure-as-code and CI/CD are mandatory. Estimated variable cost for 4 million files (≈12 TB/year) is about $462 annually plus $63 fixed, and the system meets scalability and cost goals.
AI-written summary · about
chapters (11)
- 0:00Intro
- 2:27Evolution of system deployment techniques
- 7:57Benefits of Serverless: Inherently scalable
- 14:04Limitations of Serverless: Cold start
- 19:34Use Cases
- 27:41Case Study: Proposed Implementation
- 31:21Case Study: Managing Infrastructure
- 38:49The role of DevOps engineers
- 40:02QA for Serverless
- 41:20CI/CD pipelines for Serverless • Automated CVCD pipelines are essential for Serverless applications
- 44:41Summary
related talks
42mServerless Tales from the Trenches
01h 00mServerless - reality or BS - notes from the trenches
54mServerless Architectures: no servers allowed!
47mServerless in production, an experience report
41mUnlocking Faster & Efficient Data Processing w/ Serverless
39mServerless in production, an experience report
01h 04mMigrating to Serverless - an experience report
40mThe serverless startup
