Manage Repository and Image Tag Metadata

This guide explains how to manage metadata for your repositories and image tags in the Image and Model Library. Proper metadata helps you track, organize, and govern images used for AI model serving and Postgres cluster provisioning.

Who should use this guide?

• Platform admins curating approved image sets.
• DevOps teams tagging images for specific environments (Dev, Test, Prod).
• AI/ML teams managing model lifecycle in AI Factory.
• Security & compliance owners applying ownership metadata.

What this enables

• Apply descriptive tags to repositories and image tags.
• Edit READMEs for repositories for team guidance.
• Track image ownership and promotion state.
• Provide clear context to end users in Image Library UI.

Estimated time to complete

5–10 minutes per repository/image.

Prerequisites

• You have access to Image and Model Library in Hybrid Manager or AI Factory.
• Your repository integration is already configured and image sync completed.
• You have appropriate permissions to edit metadata.

Supported Metadata Features

Managing Repository Metadata

Apply HCP Tags to a Repository

1. Navigate to Image and Model Library.
2. Click on desired Repository (e.g. postgresql, meta/llama-3.3, your-private-model-repo).
3. Click Manage Tags or Edit Tags.
4. Add HCP Tags:

• Example: approved-production, beta, team-ml, archived.

5. Save changes.

Edit Repository README

1. On the Repository page, click Edit README.
2. Enter or update README content (supports Markdown).

• Example sections:
• Purpose of repository.
• Model or DB image types contained.
• Support contacts.
• Security notes.

3. Save README. It will now display on Repository details page.

Managing Image Tag Metadata

Apply HCP Tags to an Image Tag

1. Navigate to Repository.
2. Click Image Tags tab.
3. For desired Image Tag, click Manage Tags or Edit Tags.
4. Add HCP Tags:

• Example: approved-production, testing, deprecated.

5. Save changes.

Common Tag Usage Patterns

Best Practices

• Always apply approved-production only after full validation.
• Clearly tag deprecated or archived images.
• Use team-based tags (team-ml, team-dba) to indicate ownership.
• Add README to every custom Repository — improves transparency.
• Use Image Tags + HCP Tags for precise version tracking (ex: postgresql:17.4-xxxx tagged with approved-production).

Limitations & Notes

• Applying Tags / README does not affect image pull or runtime behavior — they are for UI and governance.
• Tags must follow naming conventions established in your org.
• Currently HCP Tags do not propagate to private registry — they are stored in HCP metadata.

Example Workflow

Scenario: ML team is preparing to move a new Reranker model into production.

1. Team syncs model image nv-rerankqa-1b-v2:1.8.2 into private-model-repo.
2. After test validation:

• Apply approved-production tag to this Image Tag.
• Apply team-ml to Repository and Image Tag.
• Edit Repository README to include model version history and ownership.

3. Now AI Factory users can easily identify approved-production Reranker model.

Summary

• You can manage HCP Tags and READMEs for Repositories and Image Tags.
• This improves governance, tracking, and transparency.
• Always tag Production-approved images clearly.
• Keep READMEs updated for team guidance.

Related Links

• Deploy AI Models from Model Library
• Integrate Private Registry
• Define Repository Rules
• Model Library Explained
• Image and Model Library Explained