| name | rhoai-model-registry |
| metadata | {"author":"rhoai3-demo","version":"1.0.0","platform-family":"rhoai","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"RHOAI Platform"} |
| description | Use when documenting, reviewing, or provisioning Red Hat OpenShift AI model registries: enabling the modelregistry DataScienceCluster component, setting registriesNamespace, creating and editing model registries, choosing default non-production or external PostgreSQL/MySQL databases, configuring database CA certificates, managing registry permissions for users, groups, and project service accounts, understanding generated registry RBAC, and deleting registries without removing connected databases. Do NOT use for model catalog source governance, user model registration/version/deployment workflows, model serving runtime configuration, or live cluster changes without the OpenShift safety guard.
|
RHOAI Model Registry
Use this skill to provision and secure access to OpenShift AI model registries
for the active product baseline in docs/PLATFORM_BASELINE.md.
Source Grounding
Read references/source-capture.md before using product configuration details.
Official Red Hat documentation is product authority. This skill adapts the
official Managing model registries guide to this repo's GitOps and governance
review model.
Scope
This skill covers:
- the difference between the model catalog and model registry
- enabling the
modelregistry component on the DataScienceCluster
- setting
registriesNamespace: rhoai-model-registries
- verifying the registry namespace and operator controller deployment
- creating model registries from the dashboard
- model registry resource naming constraints
- default non-production database versus external PostgreSQL or MySQL database
selection
- external database host, port, username, password, database name, and CA
certificate options
- editing model registry name, description, and database details
- granting registry access to OpenShift users, groups, and project service
accounts
- generated registry RBAC resources and
<model-registry-name>-users groups
- deleting model registries and preserving connected databases
Use other skills for adjacent work:
rhoai-model-catalog-sources for AI hub catalog source governance
rhoai-model-catalog-workflows for catalog discovery, performance
evaluation, and starting catalog registration or deployment flows
rhoai-model-registry-workflows for registering models, versions, ModelCar
images, transfer jobs, metadata edits, deployments, archive, and restore
rhoai-model-serving-platform for serving models from registries
rhoai-dashboard-customization for dashboard visibility flags such as
disableModelRegistry
rhoai-users-groups-access for broader OpenShift AI user and group access
rhoai-certificate-management for shared CA bundle behavior
rhoai-api-tiers for modelregistry.opendatahub.io/v1beta1 support posture
Demo Policy
For this repo:
- Treat model registry as the governed metadata and lifecycle store between
experimentation and serving.
- Use the default database only for demo, evaluation, development, or testing
flows, and label it as non-production.
- For production-oriented European enterprise narratives, describe external
PostgreSQL or MySQL as the production path.
- Do not store database passwords or uploaded certificate material directly in
committed manifests.
- Prefer group-based access for repeatable demo roles; use individual users
only for narrow exceptions.
- Do not grant
system:authenticated access unless the README or operations
note explicitly explains why all cluster users should access the registry.
- Treat registry deletion as a destructive action requiring explicit approval;
connected databases are not removed by deleting the registry.
- Do not author a direct
ModelRegistry custom resource manifest unless the
exact API, fields, and support posture are verified against active Red Hat
docs or live schema.
Workflow
- Confirm the active baseline in
docs/PLATFORM_BASELINE.md.
- Read
references/source-capture.md and
references/official-doc-extraction.md.
- Confirm the
modelregistry component should be managed by OpenShift AI.
- Enable or review the
DataScienceCluster component configuration.
- Decide whether the registry uses the default non-production database or an
external PostgreSQL/MySQL database.
- For external databases, document hostname format, port, credentials source,
database name, and CA certificate handling.
- Create or edit the registry through the documented dashboard workflow unless
a schema-verified GitOps path exists.
- Grant access through users, groups, or projects and review generated RBAC.
- Verify the registry in Model registry settings or AI registry settings.
- Validate with
references/validation-checklist.md.
References
references/source-capture.md
references/official-doc-extraction.md
references/validation-checklist.md
examples/model-registry-patterns.md