Scaling Platform Engineering Without Slowing Developers Down
summary
Practical recipe: adopt CDK Terrain with TypeScript typing, treat IaC as an API, and publish modules in a catalog/IDP with versioning, tests and scaffolding. Implement a three-layer abstraction (primitives → meta features → full features) with deliberate escape hatches so 80% use the full feature while ~2–5% use primitives. Example: replace a four-day S3 request workflow with a five-line import that deploys a compliant bucket (logging, tags, alarms, least-privilege IAM) and compiles before CI.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker intro
- 1:29The slow ticket problem
- 3:56Four ways IaC breaks at scale
- 7:26Infrastructure as code as an API
- 9:13Layering modules for reuse
- 17:21Platform team responsibilities
- 19:06Testing and versioning modules
- 21:12Making modules easy to find
- 22:37AI-friendly infrastructure as code
- 24:41Continuous improvement flywheel
- 27:30Three key takeaways
- 29:20Before and after the solution
- 30:03Q&A session
related talks
29mThe Key to Faster Delivery: Collaborative Platform Strategies
41mScaling Infrastructure as Code
29mIaC Next Steps in Platform Engineering
29mThe Platform Engineering Revolution Panel: Empowering Developers Through Self-Service
40mScaling Startup for Success: Building modern Architectures and Developer Acceleration
33mPlatform engineering: Build your own PaaS
16mScaling ops: Embracing platform engineering’s potential
46mWork It Smarter, Dev It Better, Ship It Faster, Test It Stronger