SUSE AI Lifecycle Manager

SUSE AI Lifecycle Manager is a Rancher UI Extension for managing SUSE AI components across Kubernetes clusters. This extension provides a unified interface for installing, managing, and monitoring AI workloads in Rancher-managed clusters.

Note: While this extension is open source (Apache 2.0), it requires an active SUSE AI subscription to access the application catalog.

Development

Prerequisites

• Node.js 20+ and Yarn
• Access to a Rancher cluster
• Extension developer features enabled in Rancher

Setup

1. Clone and install dependencies:

   git clone <repository-url>
   cd suse-ai-lifecycle-manager
   yarn install

2. Build the extension:

   yarn build-pkg suse-ai-lifecycle-manager

3. Serve during development:

   yarn serve-pkgs

   Copy the URL shown in the terminal.

4. Load in Rancher:

   • In Rancher, go to your user profile (top right) → Preferences
   • Enable "Extension Developer Features"
   • Navigate to Extensions from the side nav
   • Click the 3 dots (top right) → Developer Load
   • Paste the URL from step 3, select "Persist"
   • Reload the page

Debug Mode

Enable debug logging in development:

NODE_ENV=development yarn build-pkg suse-ai-lifecycle-manager

Building for Production

yarn build-pkg suse-ai-lifecycle-manager --mode production

Extension Catalog Container

• The container packages the SUSE AI Lifecycle Manager (Rancher UI Extension) into a single OCI container image.
• This container is:
  • Built and published during CI
  • Stored in GitHub Container Registry (GHCR)
  • Consumed by Rancher as an extension catalog source
• The catalog container allows:
  • Versioned releases
  • Immutable distribution
  • Simple rollout via container tags

Versioning

• The catalog container tag is derived from the Git tag:

suse-ai-lifecycle-manager-<version> → ghcr.io/suse/suse-ai-lifecycle-manager:<version>

In the examples below, <version> refers to a published extension release (e.g. 0.2.0).

Available catalog image versions are published in GitHub Container Registry: https://github.com/SUSE/suse-ai-lifecycle-manager/pkgs/container/suse-ai-lifecycle-manager

Container Structure

/home/plugin-server
└── plugin-contents/
    ├── files.txt
    ├── index.yaml
    └── plugin/
        ├── index.yaml
        ├── package.json
        ├── suse-ai-lifecycle-manager
            └── suse-ai-lifecycle-manager-<version>.tgz
        └── suse-ai-lifecycle-manager-<version>
            ├── files.txt
            └── plugin/
                └── <plugin source code>

Consuming the Catalog in Rancher

• Add the catalog source in the Rancher Dashboard:
  1. Navigate to Extensions → Manage Extensions Catalog
  2. Import Extension Catalog → Use the Catalog Image Reference: ghcr.io/suse/suse-ai-lifecycle-manager:<version> → Press Load
  3. From the Extensions page, Go to Manage Repositories. Verify if the SUSE AI Rancher Extension repository has the Active state. If not, refresh the connection.
  4. Go back to Extensions and install SUSE AI Rancher Extension.
  5. Re-load Rancher Dashboard. Reload the browser to ensure the extension is loaded in the UI (ctrl+r or F5 or cmd+r).
  6. The "SUSE AI Lifecycle Manager" logo will now appear on the left panel of the Rancher Dashboard.

NOTE: Replace <version> with a tag published in GitHub Container Registry. NOTE: Newly published catalogs are not always available immediately. If the catalog does not show up after publishing, navigate to Extensions → Manage Repositories and manually refresh the repository to force a re-sync.

Extension GitHub Branch

• In addition to the OCI-based catalog container, the SUSE AI Lifecycle Manager extension can be distributed via a GitHub branch. This method hosts the extension artifacts as files within a specific branch of your repository, allowing Rancher to consume the extension directly from there.

Overview

• The extension is built into static assets (index.yaml, .tgz, etc.)
• These assets are published to a GitHub branch (example, gh-pages)
• Rancher consumes the extension catalog via the repo url and branch.

GitHub Branch Structure

• Once the artifacts are pushed to the GitHub branch, the repository will expose the extension files like so:

https://github.com/<org>/<repo>/tree/gh-pages
├── index.yaml
├── assets/
│   ├── index.yaml
│   └── suse-ai-lifecycle-manager/
│       ├── suse-ai-lifecycle-manager-<version>.tgz
│       └── ...
├── charts/
│   └── suse-ai-lifecycle-manager/
│       ├── <version>/
│       │   ├── templates/
│       │   │   ├── _helpers.tpl
│       │   │   └── cr.yaml
│       │   ├── Chart.yaml
│       │   └── values.yaml
│       └── ...
└── extensions/
    └── suse-ai-lifecycle-manager/
        ├── <version>/
        │   ├── plugin/
        │   │   └── ...
        │   └── files.txt
        └── ...

This structure mirrors the catalog format that Rancher expects.

Consuming the Extension from the GitHub Branch in Rancher

• To load the extension from a GitHub branch:
  1. Navigate to Extensions → Manage Repositories
  2. Click Create New Repository
  3. Add a Name, then select Git repository containing Helm chart or cluster template definitions
  4. Enter the Git Repo URL and the Git Branch (e.g., gh-pages)
  5. Click Create
  6. Wait until the the SUSE AI Lifecycle Manager repository has the Active state.
  7. Go back to Extensions and install SUSE AI Lifecycle Manager.
  8. Re-load Rancher Dashboard. Reload the browser to ensure the extension is loaded in the UI (ctrl+r or F5 or cmd+r).
  9. The "SUSE AI Lifecycle Manager" logo will now appear on the left panel of the Rancher Dashboard.

Contributing

When contributing to this extension:

1. Follow Standard Patterns: Use the established domain model and store patterns
2. Component Organization: Place components in appropriate directories (formatters/, validators/, pages/)
3. Type Safety: Maintain strict TypeScript usage, avoid any types
4. Internationalization: Add translation keys to l10n/en-us.json for new UI text
5. Code Quality: Run yarn lint and ensure all pre-commit hooks pass
6. Feature Flags: Use feature flags for new functionality
7. Manual Testing: Ensure all functionality works across multi-cluster scenarios

Commit Message Format

This project uses conventional commits enforced by commitlint:

type: subject

body (optional)

footer (optional)

Valid types: feat, fix, docs, style, refactor, perf, test, build, ci, chore, revert, wip, deps, security

Example:

git commit -m "feat: add multi-cluster installation support"
git commit -m "fix: resolve app installation error handling"

Troubleshooting

Common Issues

1. Extension not loading: Verify URL in developer tools console
2. Build errors: Check Node.js version compatibility (requires 20+)
3. API errors: Verify cluster permissions and connectivity
4. Linting errors: Run cd pkg/suse-ai-lifecycle-manager && yarn lint to see details