Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform
Private Foundry Hub Project - Cannot edit Prompt Flow in Foundry Portal
Christopher Wilson
0
Reputation points Microsoft Employee
## Steps to Reproduce
1. Follow steps in lab, https://microsoftlearning.github.io/mslearn-ai-studio/Instructions/03-Use-prompt-flow-chat.html. *NOTE: The lab shows public access, but you need to depoloy using private Microsoft Foundry Hub and Project.
2. Deploy AI Foundry Hub and Project with private network configuration:
- Public network access: Disabled
- Private endpoints enabled
- Private DNS zones configured
- Access via VM in same VNet
3. Assign user the following roles:
- Contributor on Resource Group
- Azure AI Developer on Hub
- Azure AI Developer on Project
- Storage Blob Data Contributor on Storage Account
- Key Vault Secrets Officer on Key Vault
4. Access AI Foundry Portal (https://ai.azure.com) from Windows VM in same VNet
5. Navigate to AI Foundry Project
6. Create a new Prompt Flow (e.g., Chat Flow). Flow will be created.
7. Attempt to click on the Prompt Flow. A modal window will appear
This flow was created by MOD Administrator, but access to the workspace and default storage account allows you to edit it. Clone it to avoid accidental overwrites
## Troubleshooting Steps Taken
### Permissions (All Successful - Issue Persists)
1. ✅ Added Azure AI Developer role to AI Foundry Project
2. ✅ Added Azure AI Developer role to AI Foundry Hub
3. ✅ Added Storage Blob Data Contributor to Storage Account
4. ✅ Added Key Vault Secrets Officer to Key Vault
5. ✅ Added Contributor role at Resource Group level
6. ✅ Verified all role assignments propagated (waited 10+ minutes)
### Browser/Session Troubleshooting
1. ✅ Cleared browser cache and cookies
2. ✅ Logged out and logged back into AI Foundry portal
3. ✅ Tried different browser (Edge, Chrome)
4. ✅ Used InPrivate/Incognito mode
5. ✅ Waited for RBAC propagation (multiple times, 5-15 minutes each)
### Resource Configuration
1. ✅ Verified AI Foundry Project system-assigned managed identity has Storage Blob Data Reader
2. ✅ Confirmed all private endpoints operational with DNS A records
3. ✅ Verified user can access storage, Key Vault (through private network)
4. ✅ Confirmed resources deployed correctly with no errors
### Comparison Testing
1. ✅ Tested same user in different tenant with PUBLIC AI Foundry - WORKS
2. ✅ Confirmed issue is specific to PRIVATE network deployment
3. ✅ Verified same user identity, same Microsoft Learn lab scenario
Foundry Tools
Foundry Tools
Locked Question. You can vote on whether it's helpful, but you can't add comments or replies or follow the question.