Managed Kubeflow Service Requests
A Service Request is a formal way to ask for support, resources, or specific actions from the Connected AI Platform team. The requests below cover Managed Kubeflow (CAIP1.5): provisioning workspaces, managing users, and connecting data.
🧠 Why use a service request? Service requests help us handle standard tasks quickly and consistently. They're ideal for non-urgent, planned needs.
Looking for a CAIP Space or an LLM API Key? Those belong to the CAIP Offboard v2: All-New CAIP stack — see CAIP Offboard v2: All-New CAIP Service Requests.
📖 Glossary / Nomenclature
CAIP Space Your umbrella project container on the Connected AI Platform. It is the central entry point for managing all CAIP services (like Kubeflow Workspaces, user/role mappings, API keys, etc.).
Managed Kubeflow Workspace A dedicated, production-ready Kubeflow environment inside your CAIP Space. It is fully operated by the CAIP team and used for ML development and orchestration.
| Concept | What it is | Where it lives |
|---|---|---|
| CAIP Space | Project container for all CAIP features | Self-Service Portal |
| Managed Kubeflow Workspace | Your ML environment (Kubeflow) | Inside a CAIP Space |
For a detailed explanation and onboarding steps, see the Onboarding Guide.
🛠️ Available Requests
🆕 New Managed Kubeflow Workspace
Request the provisioning of a new Managed Kubeflow Workspace for orchestrating machine learning workflows. Ideal for teams who want a managed environment for Kubeflow pipelines and notebooks. By submitting this request, we will create a new Managed Kubeflow Workspace for you and associate it with your existing CAIP Space. You will be added as Developer to the Cloud Environment.
You need a CAIP Space before creating a Managed Kubeflow Workspace. Please make sure you already have a CAIP Space or have recently requested the creation of a new CAIP Space.
👥 Manage CAIP Space Users
Add or remove users from your CAIP Space. Use this request to manage team membership, grant access to new colleagues, or remove users who no longer need access.
🔐 Establish CI Trust
CI Trust establishes a secure, automated connection between your GitHub repository and your Managed Kubeflow Workspace on the Connected AI Platform. This mapping enables your CI/CD systems (such as GitHub Actions) to securely log in and interact with your CAIP environments (clusters and namespaces) without the need for long-lived credentials.
By setting up CI Trust, you ensure that only trusted CI pipelines can assume roles and access resources in your CAIP Space, following best practices for security and automation.
A secure relationship between your GitHub repo and your CAIP Space will be established. With this association, your GitHub Actions pipelines can safely authenticate and perform actions in your environment using role assumptions.
🔗 Establish CDH Connection
Connect your CAIP Space to the CDH Environment for data access. Use this request to enable integration with BMW's Cloud Data Hub for data engineering and analytics.
❓ Not Sure Which One to Choose?
If your request doesn't fit one of the options above, or you just have a general question:
📬 General Inquiry / Request
Use this for any non-standard request or open question about the platform. This is the best option if you are unsure which category your request falls under or need general support.
All actions above must go through the Service Request process — please don't try to work around them via Pull Request. If your request is genuinely not covered by any Service Request above, see Requests & Self-Service for guidance and the PR-based Self-Service guides (e.g. enabling MLflow, EMR-on-EKS, Grafana, KServe, or Vector Database, custom IAM policies, CDH datasets/PII API, or Vehicle Shadow). For the general fork-and-pull-request contribution model behind this workflow, see Contributing via Pull Requests.