Azure AI Foundry
The Azure AI Foundry integration discovers agents, projects, models, access, and governed activity in your Azure subscriptions. The connector performs read-only discovery and monitoring; it does not perform Foundry write actions.
Where it shows up
- Overview › Agentic estate: the Azure AI Foundry tab under Connected platforms.
- Agents: the Azure AI Foundry source coverage chip; Foundry agents appear in the agent list.
- Graph: the Foundry sensor chip.
- Settings › Organization › Data sources: an Included in console toggle. See Data sources.
Set it up
Setup is a four-step drawer. You use a service principal that you own; no Quilr-owned Entra application or redirect URI is needed.
1. Prepare the Azure account
- In Microsoft Entra ID, create an app registration for accounts in this organization only. Record its Directory (tenant) ID and Application (client) ID.
- Under Certificates & secrets, create a client secret and copy the secret Value (not the secret ID).
- For every subscription Quilr should monitor, assign these roles to the app's service principal at subscription scope:
The drawer shows the role definition IDs and an Azure CLI template for the role assignments. Repeat the assignments for each subscription.
The Foundry User role includes broad Foundry data-plane actions and permission to list Cognitive Services account keys. Quilr uses it for read-only discovery and monitoring only.
2. Connect the service principal
Enter Connection name, Directory (tenant) ID, Application (client) ID, and Client secret value, then click Check credentials. The secret is encrypted for your Quilr tenant and is never returned.
3. Choose subscriptions
Quilr lists the Azure subscriptions it can see, with the effective access for Management, Foundry, and Log Analytics. Select the subscriptions to monitor. Check Azure access re-checks permissions. Click Save subscriptions.
4. Configure project monitoring
For each discovered project, the table shows its generation (for example foundry or foundry classic), an Inventory and a Conversations checkbox, and a status with counts of agents, deployments, and threads. Choose what to monitor and click Save monitoring. Refresh projects re-runs discovery; projects that fail show an inline error.
Check that it works
The project table shows counts of agents, deployments, and threads for each monitored project, and Foundry agents appear on Agents under the Azure AI Foundry source coverage chip and on Overview › Agentic estate.