Skip to content

Conversation

@mohamedsaif
Copy link
Contributor

This pull request introduces major enhancements to the AI infrastructure provisioning scripts, focusing on onboarding use cases for the AI Hub Gateway, expanding support for new services (AI Foundry, API Center), and improving resource configuration flexibility and security defaults. The changes include a comprehensive onboarding guide, new parameters and configuration for AI Foundry and API Center, dynamic backend/model configuration, and updates to networking and monitoring modules.

AI Hub Gateway Onboarding Documentation:

  • Added a detailed onboarding guide (guides/Citadel-Access-Contracts.md) that explains how to automate use-case onboarding for AI Hub Gateway, including visual diagrams, folder structure, parameters, outputs, customization, troubleshooting, and a practical Financial Assistant example.

Support for New Services and Resource Configuration:

  • Introduced parameters and abbreviations for API Center (apicServiceName, apicSku, abbreviation in infra/abbreviations.json) and AI Foundry (aiFoundryResourceName, enableAIFoundry, aiFoundryInstances, aiFoundryModelsConfig) to support their deployment and integration. [1] [2] [3] [4] [5]
  • Added parameters for Application Insights resources and dashboards for Foundry, enabling monitoring of new services. [1] [2]

Dynamic Model and Backend Configuration:

  • Refactored model deployment configuration: removed static openAiInstances object and replaced with dynamic arrays for AI Foundry and AI Search instances, plus a flexible aiFoundryModelsConfig array to specify model deployments per Foundry instance.
  • Added logic to dynamically generate LLM backend configuration from AI Foundry instances and models, supporting multi-model and multi-backend routing in APIM.

Networking and Security Improvements:

  • Added private DNS zone support for AI Foundry (privatelink.services.ai.azure.com) and ensured DNS modules receive tags for better resource tracking. [1] [2] [3]
  • Changed default public network access for Cosmos DB to Disabled for improved security, and updated APIM SKU default to StandardV2. [1] [2]

Monitoring and Observability Enhancements:

  • Updated monitoring module to provision Application Insights resources and dashboards for AI Foundry, ensuring observability for new workloads.

These changes collectively provide a scalable, secure, and flexible foundation for onboarding AI use cases and integrating new AI services into the APIM-based gateway architecture.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant