Architecture

This architecture illustrates the two key personas and their workflows:

• PaaS Admins configure compute profiles with hardware, GPU, and parameter options
• End Users select a pre-defined profile, enter minimal details, and deploy a fully provisioned Kubernetes cluster in a single click using Rafay-managed bare metal or VM inventory

PaaS Admin Flow

sequenceDiagram
    participant Admin as PaaS Admin
    participant Portal as PaaS Studio

    Admin->>Portal: Login to PaaS Studio
    Portal-->>Admin: Home Page
    Admin->>Portal: Navigate → Compute Profiles
    Admin->>Portal: Create or Select Existing Profile
    Admin->>Portal: Customize Input Settings (CPU/GPU, Node Count, Storage)
    Admin->>Portal: Enable "Override" for editable fields
    Admin->>Portal: Hide Advanced Configurations
    Admin->>Portal: Save & Publish Profile for End Users

End User Flow

sequenceDiagram
    participant User as End User
    participant DevHub as Developer Hub
    participant Rafay as Rafay Inventory
    participant K8s as Kubernetes Cluster

    User->>DevHub: Login to Developer Hub
    DevHub-->>User: Dashboard
    User->>DevHub: Go to Compute Profiles
    User->>DevHub: Select Shared Kubernetes Profile
    User->>DevHub: Enter Minimal Inputs (Cluster Name, Worker Node Count)
    User->>DevHub: Modify parameters (if Override enabled)
    User->>DevHub: Click "Deploy"
    DevHub->>Rafay: Provision Nodes
    Rafay->>K8s: Deploy Control Plane & Worker Nodes
    K8s->>K8s: Apply Labels, Annotations & GPU Scheduling
    K8s->>K8s: Run Provisioning Workflow (Networking, Storage)
    K8s->>K8s: Apply Blueprint (Standard/GPU-Enabled)
    K8s-->>User: Provide Kubeconfig

---