microsoft-foundry-classic

What this skill does

# Microsoft Foundry Classic Skill

This skill provides expert guidance for Microsoft Foundry Classic. Covers troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. It combines local quick-reference content with remote documentation fetching capabilities.

## How to Use This Skill

> **IMPORTANT for Agent**: Use the **Category Index** below to locate relevant sections. For categories with line ranges (e.g., `L35-L120`), use `read_file` with the specified lines. For categories with file links (e.g., `[security.md](security.md)`), use `read_file` on the linked reference file

> **IMPORTANT for Agent**: If `metadata.generated_at` is more than 3 months old, suggest the user pull the latest version from the repository. If `mcp_microsoftdocs` tools are not available, suggest the user install it: [Installation Guide](https://github.com/MicrosoftDocs/mcp/blob/main/README.md)

This skill requires **network access** to fetch documentation content:
- **Preferred**: Use `mcp_microsoftdocs:microsoft_docs_fetch` with query string `from=learn-agent-skill`. Returns Markdown.
- **Fallback**: Use `fetch_webpage` with query string `from=learn-agent-skill&accept=text/markdown`. Returns Markdown.

## Category Index

| Category | Lines | Description |
|----------|-------|-------------|
| Troubleshooting | L37-L46 | Diagnosing and fixing Foundry classic issues: prompt flow compute, deployments/monitoring, private endpoints, Azure OpenAI (incl. fine-tuning), risks & safety alerts, and known portal bugs. |
| Best Practices | L47-L61 | Best practices for designing system/safety prompts, fine-tuning and using GPT/DeepSeek models, optimizing latency/throughput, and evaluating/operating Foundry chat apps in production |
| Decision Making | L62-L90 | Guidance on choosing models, regions, deployments, billing, and lifecycle strategies for Foundry and Azure OpenAI, including migration, PTU sizing, and cost/performance tradeoffs. |
| Architecture & Design Patterns | L91-L99 | Designing multi-agent architectures, configuring Foundry Agent Service for resilience, and understanding model router behavior, failover, and disaster recovery strategies. |
| Limits & Quotas | L100-L114 | Quotas, rate limits, and capacity management for Foundry classic: agent service, models, deployments, regions, dynamic/provisioned throughput, Azure OpenAI quota, batch, and fine-tuning. |
| Security | L115-L161 | Security, privacy, and compliance for Foundry: auth/RBAC, encryption and CMK, network isolation (VNets, Private Link, perimeters), Azure Policy guardrails, content filters, and data handling for models and tools. |
| Configuration | L162-L214 | Configuring, monitoring, and evaluating Foundry Classic/Models and Azure OpenAI resources, including agents, networking, storage, compute, RAG, safety, tracing, and continuous quality monitoring. |
| Integrations & Coding Patterns | L215-L335 | Patterns and code for integrating Foundry and Azure OpenAI with tools, data, and runtimes—search/RAG, MCP/OpenAPI tools, Functions/Logic Apps, SDKs, fine-tuning, realtime audio, and external data stores. |
| Deployment | L336-L358 | Deploying Foundry hubs/models (CLI, portal, Bicep, Terraform), managed/serverless endpoints, region/feature availability, and integrating deployments/evaluations with DevOps and GitHub. |

### Troubleshooting
| Topic | URL |
|-------|-----|
| Troubleshoot Prompt Flow compute session issues in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/prompt-flow-troubleshoot |
| Troubleshoot Foundry deployments and monitoring issues | https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/troubleshoot-deploy-and-monitor |
| Troubleshoot Foundry private endpoint connection errors | https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/troubleshoot-secure-connection-project |
| Troubleshoot Azure OpenAI fine-tuning in Foundry classic | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-troubleshoot |
| Monitor and troubleshoot Risks & Safety in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/risks-safety-monitor |
| Resolve known issues in Microsoft Foundry classic portal | https://learn.microsoft.com/en-us/azure/foundry-classic/reference/foundry-known-issues |

### Best Practices
| Topic | URL |
|-------|-----|
| Deploy and use DeepSeek-R1 reasoning model in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/tutorials/get-started-deepseek-r1 |
| Design effective system messages for Azure OpenAI | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/advanced-prompt-engineering |
| Apply safety system message templates in Azure OpenAI | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/safety-system-message-templates |
| Author safety system messages for Azure OpenAI | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/system-message |
| Apply safety evaluation when fine-tuning Foundry models | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-safety-evaluation |
| Fine-tune GPT-4 vision models with image data | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-vision |
| Optimize Azure OpenAI latency and throughput in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/latency |
| Apply best practices for Azure OpenAI On Your Data | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/on-your-data-best-practices |
| Optimize Azure OpenAI predicted outputs for latency | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/predicted-outputs |
| Operate Foundry provisioned throughput in production | https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/provisioned-get-started |
| Evaluate and improve Foundry-based chat apps with SDK | https://learn.microsoft.com/en-us/azure/foundry-classic/tutorials/copilot-sdk-evaluate |

### Decision Making
| Topic | URL |
|-------|-----|
| Select Azure OpenAI models and regions for Foundry Agents | https://learn.microsoft.com/en-us/azure/foundry-classic/agents/concepts/model-region-support |
| Decide when and how to fine-tune models in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/fine-tuning-overview |
| Compare Foundry models using benchmarks and leaderboards | https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/model-benchmarks |
| Manage lifecycle of managed compute Foundry models | https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/model-retirement-managed-compute |
| Choose the right Azure resource type for Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/resource-types |
| Choose Microsoft Foundry deployment types | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/deployment-types |
| Plan and manage model versioning in Foundry Models | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/model-versions |
| Plan and manage model versioning in Foundry Models | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/model-versions |
| Select Azure regions and deployments for Foundry models | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/models-sold-directly-by-azure-region-availability |
| Choose between GPT-5 and GPT-4.1 in Foundry | https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/model-choice-guide |
| Compare models with Foundry leaderboards | https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/benchmark-model-in-catalog |
| Plan and manage costs for Microsoft Foundry hubs | https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/costs-plan-manage |
| Migrate from hub-base