Building a Platform Engineering API Layer with KCP
summary
KCP claims to be a horizontally scalable control plane that presents isolated workspaces spin up in milliseconds and returns kubeconfigs with workspace paths so kubectl just works; demos create workspaces live, show API exports and API bindings to publish CRD-like schemas, and expose virtual workspaces feed existing controllers so controllers outside KCP see tenant objects. It deliberately has no workload scheduling dependency and can run as a standalone web service, and it’s a CNCF Sandbox project that mostly works as demonstrated.
AI-written summary · about
chapters (10)
- 0:00Platform engineering and KCP
- 3:37Focus on the API layer
- 7:30What is KCP
- 10:58KCP workspaces and isolation
- 14:14API exports and bindings
- 17:46Virtual workspaces for APIs
- 20:15Universal API layers
- 28:42Why KCP for APIs
- 30:09Q&A: KCP vs VCluster
- 33:21Closing remarks
related talks
27mBuilding a Platform Engineering API Layer with kcp
32mBuilding a Platform Engineering API Layer with kcp
05mkcp: Scaling The Kubernetes Control Plane For The Multi-Clust...
01h 15mTutorial: Exploring Multi-Tenant Kubernetes APIs and Controllers With Kcp
22mTaming the VM Sprawl: KCP Meets KubeVirt
26mKubernetes - The Universal Control Plane
41mKcp: Towards 1,000,000 Clusters, Name^WWorkspaced CRDs
06mKubernetes as the Control Plane for the Hybrid Cloud