Scaling Infrastructure as Code
summary
Argues for splitting monolithic IaC into independently deployable stacks with clear interfaces, pipelines, and progressive tests; shows a real case where terraform apply took two hours and explains why modules alone don’t solve blast-radius and change-speed problems. Uses concrete tools and patterns: Terraform/CDK/CloudFormation, Helm, LocalStack, Prometheus, S3/Route53/ACM, versioned stack artifacts, CI pipelines and test fixtures. Recommends split into independent stacks, define stack-boundary interfaces and automated promotion.
AI-written summary · about
chapters (9)
- 0:00Intro
- 1:24What the organization needs its infrastructure to support
- 4:58Challenges we have with scaling infrastructure
- 14:23A centralized platform architecture
- 15:57A collaborative platform architecture
- 22:57An infrastructure stack is an architectural quantum
- 26:49Complexity moves to the boundaries between stacks
- 28:52Aligning boundaries of change
- 31:49Test stack code in isolation to ensure loose coupling
related talks
33mScaling Infrastructure as Code
38mApplication-Driven Infrastructure Deployment
44mTaming the cloud sprawl
46mImplementing Infrastructure as Code
27mApplication-Driven Infrastructure
48mMoving from spaghetti infrastructure to composable cloud environments
22mLessons Learned from Scaling Infrastructure as Code
46mThink of your infrastructure as code