How to create a serverless center of excellence
summary
Capital One describes creating an enterprise serverless COE that governs thousands of accounts and tens of thousands of Lambdas, enforces IAM, naming and tagging standards, and uses Org SCPs and Cloud Custodian to auto-detect and remediate noncompliance. Tooling includes the SAM CLI for local debugging, a custom CI/CD pipeline, and Lambda Power Tuner to pick memory (512–1GB typical) and cut cost; SnapStart slashes Java cold starts by 90%+.
AI-written summary · about
chapters (25)
- 0:00Introduction
- 3:16Why serverless
- 4:02Preserverless world
- 6:27Serverless world
- 10:16Everything at the bottom
- 11:27Summary
- 12:36Chapter 2 Lessons
- 14:56Chapter 4 Center of Excellence
- 18:42Chapter 6 Lambda
- 22:02Chapter 8 Learning Curve
- 22:27Chapter 9 Debugging
- 23:40Chapter 10 Sam
- 26:19Chapter 12 Demo
- 26:31Chapter 13 Demo
- 29:22Lambda Logs
- 33:23Concurrency
- 35:19Concurrency Formula
- 37:01Set Development Standards
- 38:58Standardize Tags
- 39:59IAM Policies
- 42:39Multiple Accounts
- 43:50Power Tools
- 44:22Lambda Pricing
- 45:49Lambda Power Tuner
- 48:23Things not to do
related talks
32mImplementing a serverless-first strategy in an enterprise
01h 01mServerless at scale
51mTaking Serverless to the Next Level
45mPatterns & practices for developing enterprise-scale serverless applications
41mYou Don’t Want to Build a Serverless Monolith
46mServerless Architecture Patterns and Best Practices
17mGoing Full Silo - Using Serverless to Build for Dev Simplicity
41mCI/CD in the Serverless Era