Scaling Infrastructure as Code
summary
Argues that infrastructure-as-code silos—Terraform, Ansible, Pulumi, CloudFormation or ad hoc Python—create snowflaked projects that don’t scale: example of teams copying region configs and a single repo where terraform apply takes two hours. Shows a working pattern: break monolith projects into independently deployable stacks (S3 website bucket, IAM roles, Route53/CDN), inject dependencies via parameters, test stacks in isolation with fixtures, and push via pipelines so teams deliver changes without heavy coordination. Recommends domain-based platform teams, developer portals and versioned infra code.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:09Business strategy and growth
- 6:02Challenges in infrastructure code
- 9:12Collaborative platform design
- 13:48Empowering vs disempowering platforms
- 17:38Scaffolding for infrastructure
- 19:27Design principles for scaling
- 23:59Breaking monolithic stacks
- 32:47Summary and Q&A
related talks
41mScaling Infrastructure as Code
46mImplementing Infrastructure as Code
54mInfrastructure as code (anti) Patterns
27mApplication-Driven Infrastructure
42mInfrastructure as Code - Lessons learned from Dev to Ops
46mThink of your infrastructure as code
23mUnleashing Potential Across Teams: Infrastructure as Code
44mTaming the cloud sprawl