Simplifying API Review - A Tool Based Approach for Kubernetes and Beyond
summary
Kubernetes API best practices get encoded into tools so reviewers stop policing mechanical issues and focus on user-facing design. A Go plugin kube-api-linter already implements about 28 rules (comments, list types, serialization checks) and is used in projects like Cluster API and Operator Lifecycle Manager; a companion tool crdify / certify runs as a standalone binary to catch schema and versioning breakage (17 rules today) and flag incompatible validation or roundtrip data loss in CI.
AI-written summary · about
chapters (11)
- 0:00API review and tooling goals
- 1:46Why API review is complex
- 3:29Key considerations in API review
- 5:32API implementation and schema
- 8:01Serialization challenges in API design
- 11:10Kube API Lint for static analysis
- 13:37Schema changes and compatibility
- 17:29Versioning rules and data loss
- 18:31Kubernetes certify for schema validation
- 20:18Project goals and future work
- 21:21Q&A
related talks
23mAPI Design Reviewer's Starter Set
36mSIG API Machinery Deep Dive
28mKubernetes CRD Design for the Long Haul: Tips, Tricks, and Lessons Learned
26mAPI Governance: The Latest Trends Making Teams Love API Standardization.
24mThe Augmented API Design Reviewer
36mOpenAPI Specs – Towards Native User Experience of CRDs
14mHow to build better and consistent APIs at scale
25mAutomating API Governance