v3.0.0

Adds end-to-end Foundry IQ retrieval support to GPT-RAG.

This release updates GPT-RAG configuration, deployment pins, and documentation so operators can deploy GPT-RAG with Foundry IQ as the retrieval backend.

What changed

• Updated GPT-RAG manifest pins for the v3.0.0 release.
• Pinned AI Landing Zone Bicep module to v2.1.0.
• Pinned gpt-rag-orchestrator to v2.9.0.
• Added post-provision and App Configuration defaults for Foundry IQ deployments.
• Aligned generated Knowledge Base, Knowledge Source, and Azure AI Search endpoint names.
• Documented Foundry IQ deployment patterns, authorization behavior, migration, rollback, billing, limitations, and troubleshooting.

Why it matters

GPT-RAG can now be deployed with either Azure AI Search retrieval or Foundry IQ retrieval using a documented, repeatable release path.

Foundry IQ deployments support:

• Pattern A: native permissions with OBO.
• Pattern B: custom security fields with filterAddOn.

Component versions

Component	Version
gpt-rag-ui	v2.3.13
gpt-rag-orchestrator	v2.9.0
gpt-rag-ingestion	v2.4.13
bicep-ptn-aiml-landing-zone	v2.1.0

Validation

Validated in a live Azure environment:

• Provision and deploy completed successfully.
• Foundry IQ direct retrieval returned HTTP 200.
• Orchestrator image smoke test returned HTTP 200.
• Logs confirmed retrieval through Foundry IQ.
• Frontend health, ingestion docs, and orchestrator docs endpoints returned HTTP 200.

References

• Tracking issue: #526
• Manifest/config PR: #528
• Docs PR: #529
• Orchestrator release: https://github.com/Azure/gpt-rag-orchestrator/releases/tag/v2.9.0
• AI Landing Zone Bicep release: https://github.com/Azure/bicep-ptn-aiml-landing-zone/releases/tag/v2.1.0