Lightning Talk: Kubeflow Profiles Automation for Declarative Us...
summary
Canonical builds an operator that continuously syncs a single source of truth (OIDC or a GitHub YAML repo) to Kubeflow Profiles, RoleBindings and authorization policies, using an IDP parser to emit a PMR (Profile Manage Representation) and an operator loop that removes or creates bindings. A separate GitHub Profiles Automator deployable via Juju watches commits, updates clusters, and exposes an admin delete stale profiles action; work runs today and Entra ID support is next.
AI-written summary · about
chapters (6)
- 0:00Kubeflow profiles automation intro
- 1:23User management challenges
- 3:31Proposed solution
- 5:41GitHub declarative profiles
- 7:26Implementation options
- 8:29Future plans
related talks
34mKubeflow: Multi-Tenant, Self-Serve, Accelerated Platform for Practitioners
31mKubeflow 2026 State of the Union: Orchestrating the Full ML Lifecycle
05mCI/CD Pipelines and Machine Learning
24mThe Good, the Bad, and the Missing Parts of Kubeflow
14mKubeflow: Where ML Meets Kubernetes
05mProject Lightning Talk: Transforming your Application Behavior to Kubernetes Objects
32mKubeflow Deep Dive
14mLightning Talk: Leveraging Kubeflow for Next-Generation AI Healthcare Solutions