Multi-tenancy in Kubernetes
summary
Argues that sharing one Kubernetes control plane plus a large cluster of nodes saves cost and increases delivery velocity by avoiding cluster sprawl, and shows the concrete controls needed: namespaces as virtual clusters, RBAC, network policies, pod sandboxes such as gVisor, and resource quotas plus taints/tolerations and QoS to prevent noisy neighbors. Soft multi-tenancy runs in production at many companies; hard multi-tenancy (zero trust) is not yet widely used and requires ongoing API‑server fairness and isolation work.
AI-written summary · about
chapters (25)
- 0:00Intro
- 1:18Who am I
- 1:36Why Multitenancy
- 2:18How Kubernetes works
- 2:48What happens when you have multiple users
- 3:30Sprawl of clusters
- 4:07How much does it cost
- 5:55Operational overhead
- 7:06Separate clusters
- 7:43Multitenancy
- 9:14Multitenancy in Kubernetes
- 10:38Cluster User
- 13:00Hard Multitenancy
- 13:40Security Isolation
- 15:03Custom Controllers
- 18:13Primitives
- 19:30Access Control
- 21:30Secrets
- 23:21Network Policies
- 24:59Custom Resource Definitions
- 25:45Sandboxes
- 28:12Fair Shearing
- 31:30Influence the Scheduler
- 33:30Control Plane
- 38:26Summary
related talks
37mMulti-Tenancy in Kubernetes: Best Practices Today, and Future Directions
24mMulti-Tenancy: Tips, Tricks, Tools And Tests
40mPanel Discussion: Multi-Tenancy in Kubernetes: Current State and Future Roadmap
26mThe Future of Multi-Tenancy in Kubernetes
05mNamespace Multi-Tenancy, But All The Problems Related To It
51mMulti-tenant Kubernetes
51mMultitenant Kubernetes: squeezing out maximum efficiency
33mKubernetes Multi-tenancy: Principles and Practices for Large Scale Internal Deve...