Add AI Gateway Foundry Models lab

[Alex Vieira (vieiraae)]

Summary

Adds a new lab — AI Gateway Foundry Models — demonstrating how to publish and govern Microsoft Foundry models through the AI Gateway tier of Azure API Management (public preview).

What's included

• labs/aigw-foundry-models/ — notebook, main.bicep, README, and clean-up notebook
• images/aigw-foundry-models.gif — lab flow diagram
• modules/ai-gateway/v1/ — new reusable modules (AI Gateway, model providers, role assignments)
• Updates to shared modules: cognitive-services/v3 (deployments.bicep, foundry.bicep), monitor/v1/appinsights.bicep, operational-insights/v1/workspaces.bicep

The lab covers

• Publishing Foundry models behind a single, stable gateway endpoint
• Calling models via the OpenAI Chat Completions (structured output) and Responses APIs
• Streaming responses with time-to-first-token (TTFT)
• Authentication checks (missing/invalid keys rejected with HTTP 401)
• The token rate-limit policy demonstrated and validated as a leaky bucket, with a visualization
• Token-usage analysis via managed Prometheus, broken down by model and subscription

The AI Gateway tier is in public preview (East US 2 and Sweden Central). See the announcement.

[github-actions]

📋 Labs Config Preview

The following labs will be updated in docs/labs-config.json when this PR is merged:

• aigw-foundry-models

The config will be updated on both main and gh-pages branches.