🔐 How to Securely Use Azure Key Vault Secrets in Microsoft Fabric Notebooks

If you're working with sensitive values like API keys (for OpenAI, Azure, or other services) in Microsoft Fabric, you should never hardcode them in your notebook.

This post shows how to securely retrieve secrets using Azure Key Vault and Fabric’s built-in utility:
notebookutils.credentials.getSecret()

---

✅ Step 1: Create a Secret in Azure Key Vault

1. Go to the Azure Portal

2. Create a Key Vault or use an existing one

3. In the Key Vault, go to Secrets

4. Click + Generate/Import

   • Name it something like openaiapikey

   • Paste your API key into the value field

   • Click Create

---

✅ Step 2: Assign Access Using RBAC

Fabric uses role-based access control (RBAC) to allow secure access to Key Vault secrets.

1. In the Key Vault, go to Access control (IAM)

2. Click + Add → Add role assignment

3. Select:

   • Role: Key Vault Secrets User

   • Assign access to: Managed identity

4. In the identity search, locate your Fabric workspace's identity

   • It may appear as Microsoft Fabric - <workspace name>

   • If not visible, refer to this YouTube walkthrough

---

✅ Step 3: Use Secrets in Fabric Notebook

In your Fabric notebook:

import openai
import pandas as pd

# Use Fabric's notebookutils to pull your Key Vault secret
openai.api_key = notebookutils.credentials.getSecret(
    'https://openaisqlbits.vault.azure.net/',  # your Key Vault URI
    'openaiapikey'  # your secret name
)

openai.api_type = "azure"
openai.api_base = "https://openaisuebayes.openai.azure.com/"
openai.api_version = "2023-07-01-preview"
DEPLOYMENT_NAME = "gpt-35-turbo"

def summarise_review(text):
    response = openai.ChatCompletion.create(
        engine=DEPLOYMENT_NAME,
        messages=[{"role": "user", "content": f"Summarise this review: {text}"}]
    )
    return response['choices'][0]['message']['content']

✅ You’ve now fully secured your API key — no need to store it in your code or environment variables.

---

🧠 Why This Matters

• No more exposing secrets in .ipynb files or GitHub

• Fabric notebooks remain clean, repeatable, and safe

• Supports real-world production use (like securely connecting to OpenAI, Azure services, REST APIs)

---

📚 Resources

• 🔗 Microsoft Docs: Notebook Utilities for Fabric

• 🎥 YouTube Walkthrough: Using Azure Key Vault secrets in Fabric

---

✅ Next up: I’ll show how to track OpenAI API costs + token usage so you can see exactly what each analysis is costing 💷