Configure OpenAI via Microsoft Foundry for Cribl AI
Use this topic to connect OpenAI models via Microsoft Foundry as a Bring Your Own Model (BYOM) for Cribl AI. This page covers:
- What you need from the Microsoft Foundry/OpenAI side.
- How to fill in each step of the Cribl configuration wizard.
For the generic wizard flow, prerequisites, and how to edit or delete a configured provider, see Configure Custom AI Providers.
Prerequisites
In addition to the general prerequisites, you need:
- A Microsoft Foundry/OpenAI deployment that your Organization manages.
- Deployed OpenAI models that can accept the model IDs Cribl sends (see Required Model IDs below).
- A Deployment URL for that OpenAI deployment (for example, an HTTPS endpoint exposed by Foundry or your internal gateway).
- An API key that your Organization exposes for OpenAI via Microsoft Foundry access (often via an internal API gateway or service).
- On-prem only: Network connectivity from the Leader to the endpoint your Organization uses for OpenAI via Microsoft Foundry (public endpoint, private link, or corporate proxy).
Required Model IDs
Cribl AI agents use one model per tier. The wizard pre-fills suggested defaults and restricts selection to supported models.
Suggested defaults:
| Tier | Model | Model ID |
|---|---|---|
| Small | GPT-5.4 Mini | gpt-5.4-mini |
| Frontier | GPT-5.4 | gpt-5.4 |
| Reasoning | GPT-5.5 | gpt-5.5 |
All supported models (available in the wizard dropdown):
| Model | Model ID | Available Tiers |
|---|---|---|
| GPT-4.1 Nano | gpt-4.1-nano | Small |
| GPT-4.1 Mini | gpt-4.1-mini | Small, Frontier |
| GPT-4.1 | gpt-4.1 | Small, Frontier, Reasoning |
| GPT-4o | gpt-4o | Small, Frontier |
| GPT-5 Nano | gpt-5-nano | Small |
| GPT-5 Mini | gpt-5-mini | Small, Frontier |
| GPT-5 | gpt-5 | Frontier, Reasoning |
| GPT-5 Pro | gpt-5-pro | Frontier, Reasoning |
| GPT-5.1 | gpt-5.1 | Frontier, Reasoning |
| GPT-5.2 | gpt-5.2 | Reasoning |
| GPT-5.4 Nano | gpt-5.4-nano | Small |
| GPT-5.4 Mini | gpt-5.4-mini | Small, Frontier |
| GPT-5.4 | gpt-5.4 | Frontier, Reasoning |
| GPT-5.4 Pro | gpt-5.4-pro | Frontier, Reasoning |
| GPT-5.5 | gpt-5.5 | Reasoning |
| o3-mini | o3-mini | Small, Frontier |
| o3 | o3 | Reasoning |
| o3-pro | o3-pro | Reasoning |
| o4-mini | o4-mini | Frontier, Reasoning |
The wizard dropdown for each tier only shows models that are eligible for that tier. For example, GPT-5.5 appears only in the Reasoning dropdown.
Make sure each model you assign is deployed in your Microsoft Foundry environment and reachable through your Deployment URL before saving. If you need to serve multiple Cribl model IDs from a smaller set of deployments, you have two options:
- Proxy remapping: Use a proxy (such as LiteLLM) to map incoming model names to your available deployments.
- API gateway aliasing: Configure your API gateway to translate model IDs before forwarding to your backend.
For details on model categories and token limits, see Model IDs and Feature Routing.
Step 1: Select the Provider Type
Navigate to your AI Settings and open the configuration wizard:
- Select Use Custom AI Provider (or Try it!).
- Select the Microsoft Foundry tile. Selecting the tile advances automatically to step 2.
- Select Continue if the wizard does not advance automatically.
Step 2: Enter Connection Details
Provide the credentials Cribl needs to reach your Foundry deployment.
ID: Enter a short, unique identifier (for example,
openai-foundry-prod). Avoid putting secrets or keys in this field.Description: Enter a human-readable label for the provider card (for example,
OpenAI via Microsoft Foundry - production). This field is optional.Deployment URL: Enter the base HTTPS endpoint URL for your Microsoft Foundry deployment. The URL must end in
/openai/v1. For example:https://<resource-name>.openai.azure.com/openai/v1
v1 API required. Cribl BYOM for OpenAI via Microsoft Foundry uses the Microsoft Foundry v1 OpenAI-compatible API at the
/openai/v1endpoint path. The pre-v1 Azure OpenAI API (the one that requires anapi-versionquery parameter) is not supported, because Microsoft has deprecated it. Before saving the configuration, make sure that:
- Your Microsoft Foundry deployment has the v1 API enabled.
- The Deployment URL you enter in the modal ends with
/openai/v1.For background on the available Foundry endpoints, see Endpoints for Microsoft Foundry Models.
Include the
https://prefix and the/openai/v1path. Do not append additional path suffixes such as/chat/completionsafter/openai/v1. Cribl appends those automatically.
- API Key: Paste the API key or bearer token provided by your platform team.
Depending on your environment setup, this key is one of the following:
- If you are using Azure OpenAI Service directly, this is the key found on the Keys and Endpoint tab in the Azure Portal.
- If using a gateway, use the token generated by your internal API management layer.
If you are unsure which key to use, ask the team that manages AI providers and Microsoft Foundry access to issue or confirm the correct key for your deployment.
Select Continue.
Step 3: Assign Models to Tiers and Test
Assign models to each tier. For Microsoft Foundry, the wizard presents a dropdown of supported model IDs with the suggested defaults pre-filled.
The Reasoning tier supports multiple model IDs, with the first model in the list serving as the system default. While most features exclusively use this first model, Cribl Search investigations allow you to switch between any listed models via a dropdown menu.
- For Small and Frontier, select a model from the dropdown or accept the suggested default.
- For Reasoning, accept the suggested default or add additional model IDs.
- Select Test and Save.
Cribl tests each assigned model by sending a request through your configured endpoint:
- A results panel shows how many models passed and failed.
- If any test fails, verify that the Deployment URL is correct, the API key is valid and active, and the listed model IDs are deployed and accessible. Select Retry to test again.
- When all models pass, the configuration is saved automatically.
After you save:
- A Custom AI provider card appears at the top of AI Settings, showing your ID, description, OpenAI via Microsoft Foundry as the provider type, and the number of configured models.
- Supported Cribl AI capabilities in that Workspace begin using OpenAI via Microsoft Foundry as their AI backend.
Verify Cribl AI Behavior
To confirm that OpenAI via Microsoft Foundry is correctly configured:
- In AI Settings, confirm that the Custom AI provider card lists OpenAI via Microsoft Foundry.
- Use a supported Cribl AI capability (for example, the KQL assistant or Git commit message suggestions).
- Verify that:
- Requests succeed without provider-related errors.
- Latency and behavior align with your expectations for the OpenAI models and deployment you configured.
If you see failures or unexpected behavior:
- Select the gear (⚙️) on the provider card and select Edit to reopen the wizard, then use Test and Save to review error details.
- Confirm with your platform/AI team that:
- The endpoint URL and key are valid and not rate-limited.
- Your OpenAI deployments are enabled and accessible behind the Foundry or gateway endpoint.
- For on-prem deployments: Network rules allow the Leader to reach the Foundry or gateway endpoint.
Change or Stop Using OpenAI via Microsoft Foundry
- To update the provider (for example, after a key rotation or endpoint change), see Edit an Existing Custom AI Provider.
- To stop using OpenAI via Microsoft Foundry, see Delete a Custom AI Provider.
- To use this provider again after deletion, follow the steps in this topic to reopen the wizard and reconfigure.