Skip to content

Docs · Updated October 2026

Deploy FabriBuddy

About 15 minutes of work, plus waiting for the first scan. You import one notebook and run it. It creates everything else, then prints the commands that add the app.

Before you start

You need Why
A workspace on a Fabric capacity (F SKU or trial), where you're Admin FabriBuddy's items live there. Fabric Apps don't run on Pro or PPU. A new workspace keeps sharing simple.
A service principal, with its secret in Azure Key Vault The scan reads your tenant as this principal. Set it up →
Key Vault Secrets User on that vault, for you The scan reads the secret as whoever runs the pipeline, or owns its schedule.
Fabric Apps (preview) turned on Admin portal → Tenant settings → Fabric Apps (preview). Needed for the app only.
Node.js 20 or later on your computer To add the app, once.

1. Run the deploy notebook

  1. Download scripts/Deploy_FabriBuddy.ipynb.
  2. In your workspace: Import → Notebook → From this computer, and choose the file.
  3. Open Deploy_FabriBuddy. Check that the language at the top is Python (not PySpark).

  4. Fill in the first code cell:

    KEY_VAULT_URL = "https://<your-vault>.vault.azure.net/"
SP_TENANT_ID = "<directory (tenant) id>"
SP_CLIENT_ID = "<application (client) id>"
SP_SECRET_NAME = "fabribuddy-sp-secret"   # the secret's name in Key Vault

    The other settings are optional:

    Setting Default
    WORKSPACES "" Workspace IDs to scan, comma-separated. Empty means every workspace.
    INCLUDE_PERSONAL False Report Audience: include people's personal My workspace.
    TENANT_NAME "" How the app names your tenant, for example "Contoso".
    SCHEDULE_DAILY_AT_UTC "05:00" The daily scan time, in UTC. "" means no schedule.
    RUN_AFTER_DEPLOY True Start the first scan when deployment finishes.

  5. Choose Run all.

The notebook prints each step:

FabriBuddy 0.1.0 from zan-tavcar/msfabric-fabribuddy@main
✓ FabriBuddy_DB
✓ FabriBuddy 0.1.0 stored in FabriBuddy_DB
✓ FabriBuddy_Scan.Notebook created
✓ Check_FabriBuddy_Version.Notebook created
✓ Load_FabriBuddy_E2E.DataPipeline created
✓ Scan: Load_FabriBuddy_E2E, FabriBuddy_Scan
✓ Storage: FabriBuddy_DB
✓ Maintenance: Check_FabriBuddy_Version, Deploy_FabriBuddy
✓ Schedule: Load_FabriBuddy_E2E daily at 05:00 UTC
✓ First scan started: open Load_FabriBuddy_E2E → View run history (10–30 minutes).

Your workspace now has three folders:

Folder Items
Scan Load_FabriBuddy_E2E (pipeline, scheduled) and FabriBuddy_Scan (notebook)
Storage FabriBuddy_DB (SQL database)
Maintenance Check_FabriBuddy_Version and Deploy_FabriBuddy (notebooks)

Try it without a service principal

To look at the app before setting anything up, deploy with the service principal left empty and RUN_AFTER_DEPLOY = False. Then run Load_FabriBuddy_E2E with the parameter demo set to true. It loads the fictional Contoso tenant instead of scanning yours.

2. Add the app

The deploy notebook's last cell prints these commands, with your workspace and database IDs already in them. Run them on your computer:

git clone https://github.com/zan-tavcar/msfabric-fabribuddy.git
cd msfabric-fabribuddy/app
npm install
npx rayfin login                         # sign in with your Fabric account
npm run connect -- <workspace-id> <sql-database-id>
npx rayfin up --workspace-id <workspace-id>
  • npm run connect points the app at your FabriBuddy_DB.
  • npx rayfin up creates the fabribuddy app in your workspace, builds it and uploads it. It takes a minute or two and prints the app's address at the end.

More on the app, including sharing, in The app.

3. Open it

Open fabribuddy in the workspace. Fabric signs you in. Until the first scan has finished, the pages say Not in the last scan; reload once it has.

Next