SIG API Machinery in the Era of AI: Updates
summary
Kubernetes API machinery in 1.36 centers on operational stability: mutating admission policy (GA) and declarative validation via struct tags to generate OpenAPI schemas, watch sharding so controllers claim keyspace shards and detect stale caches, and rollout safety with mix-version proxy, two‑step upgrades, version emulation and rollback; removed an unmaintained go protobuf fork. 1.37 focuses on websocket and watch cache/storage migrations, concurrent watches and revitalizing the Python client with async informers, streaming and performance for AI workloads. AI pressures appear as workload scale, API clients (agents), and code contributors.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:03What is API Machinery
- 3:59API Machinery's role
- 4:45Kubernetes 1.36 highlights
- 5:07Mutating Admission Policy GA
- 7:16Declarative validation
- 9:25Scaling control plane
- 11:17Operability improvements
- 13:26Technical debt cleanup
- 14:27Kubernetes 1.37 expectations
- 15:31Python client for AI
- 17:40AI impact on control plane
- 20:32API performance for AI
- 21:36OpenAPI spec generation
- 22:40AI code policy
- 24:30Contributing to API Machinery
- 28:00Q&A
related talks
30mSIG API Machinery: Project Updates and Release Planning
38mDeep Dive: API Machinery SIG
33mMeet the Gears Behind Kubernetes APIs: Introduction to SIG API-Machinery
36mSIG API Machinery Deep Dive
28mAdvanced API Machinery Topics: Aggregated API Servers and OpenAPI v3
22mAI in Kubernetes The Promise
43mKeynote: Building AI in the Open: How the Cloud Native Community Shapes What Comes Next
28mAI and the state of APIs in 2026