Serverless at scale
summary
Capital One describes concrete serverless patterns and got measurable results: static website hosting using ALB+Lambda (Go) serves 20–80 KB pages in ~20–40 ms, uses gzip to push a 14 MB JS bundle, and adds memcache to cut S3 calls. Centralized logging uses cross-account subscriptions into Kinesis (one worker per shard), Glue+Athena for s3 “sandwich” analytics, Step Functions with DynamoDB (watch the 32,768‑char payload limit), Lambda@Edge control-plane limits, multi‑region S3 via Lambda multi‑writes, and careful handling of hot lambda shared memory and circuit breakers.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 2:54Serverless Architecture
- 5:20Lambda Functions
- 11:53API Server
- 13:36EventDriven Lambda
- 18:41Glue and Athena
- 20:25Rules Engine
- 24:52Service Patterns
- 31:47Lessons Learned
- 38:47Private API Endpoints
- 41:07Show of Hands
- 42:22Build Time
- 47:33Questions
- 50:47Deployment Architecture
- 53:03Infrastructure as Code
- 55:22Performance Testing
- 57:52Security
related talks
49mServerless Architectural Patterns and Best Practices
47mAdvanced Serverless Architectural Patterns on AWS
54mServerless Architectures: no servers allowed!
01h 00mServerless - reality or BS - notes from the trenches
31mAccelerate business innovation with serverless applications
42mNo Server Is Better Than No Server: Building Blocks for Serverless Architectures
47mServerless in production, an experience report
54mServerless beyond Functions