rhoai-feature-store

name	rhoai-feature-store
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 rebuilding Red Hat OpenShift AI Feature Store capabilities from the official Feature Store guide: machine learning feature concepts, Feast Operator enablement, DataScienceCluster feastoperator state, FeatureStore custom resources, registry/offline/online stores, UI visibility, Kubernetes or OIDC authorization, workbench client configuration, feature definitions, entities, feature views, historical and online retrieval, Ray and Spark compute engines, monitoring, scaling, and the feast CLI. Do NOT use for model serving, AI Pipelines, MLflow, model registry, or live cluster deployment; use the matching rhoai-* or env-* skill instead.

RHOAI Feature Store

Use this skill to ground Feature Store work in the active Red Hat OpenShift AI baseline recorded in docs/PLATFORM_BASELINE.md.

Source Grounding

Read references/source-capture.md before using product configuration details. Official Red Hat documentation is product authority. Upstream Feast material linked from the Red Hat guide is supplemental only.

Scope

This skill covers Feature Store as a RHOAI platform capability:

• enabling the feastoperator component in the DataScienceCluster
• creating FeatureStore resources in data science projects
• configuring registry, offline store, online store, UI, auth, persistence, scaling, monitoring, and disconnected behavior
• defining data sources, entities, feature views, and supported data types
• connecting workbenches to Feature Store client configuration
• using Ray, Spark, and local compute engines for materialization and historical retrieval
• using feast CLI commands in workbenches or controlled automation

Use other skills for adjacent areas:

• rhoai-workbenches-custom-images for workbench image and Notebook behavior
• rhoai-kueue-workload-management and rhoai-distributed-workload-operations for Ray/Kueue operations
• rhoai-kfp-pipeline-authoring for Kubeflow Pipelines
• rhoai-mlflow for MLflow tracking, SDK authentication, experiments, and artifact storage behavior
• env-deploy-and-evaluate for live cluster deployment flow

Demo Policy

For this repo:

• Treat Feature Store as an optional RHOAI capability until a demo step introduces it with docs, GitOps, scripts, and validation together.
• Use GitOps for DataScienceCluster and FeatureStore resources once active GitOps folders are recreated.
• Keep database credentials, OIDC client details, object-storage credentials, and registry connection strings out of Git.
• Prefer project-scoped FeatureStore resources unless a future architecture explicitly introduces a shared registry namespace.
• Do not scale Feature Store replicas above one while using file-backed SQLite, DuckDB, or local registry storage.
• Use database-backed persistence before enabling static replicas, HPA, or production-style availability.
• Validate FeatureStore schema on the target cluster before adding fields not already captured from the official guide.

Workflow

1. Confirm the active baseline in docs/PLATFORM_BASELINE.md.
2. Read references/source-capture.md and references/official-doc-extraction.md.
3. Confirm OpenShift AI is installed and the target project exists.
4. Enable Feature Store by setting DataScienceCluster.spec.components.feastoperator.managementState to Managed.
5. Create a project-scoped FeatureStore custom resource with apiVersion: feast.dev/v1.
6. Add feature-store-ui: enabled only when the OpenShift AI dashboard should expose the Feature Store UI or workbench selection flow.
7. Configure registry, online store, offline store, auth, and persistence from official examples or verified cluster schema.
8. Add workbench access and RBAC/OIDC behavior before documenting user access.
9. Use compute engines only after Ray, Spark, or local execution prerequisites are documented.
10. Validate with references/validation-checklist.md.

References

• references/source-capture.md
• references/official-doc-extraction.md
• references/validation-checklist.md
• examples/feature-store-patterns.md