Audit a Power BI on Databricks setup against the Databricks Power BI on Databricks Best Practices Cheat Sheet. Produces a single Markdown or HTML report you can share with your team.

42 rules across four phases:

See docs/superpowers/specs/2026-05-01-powerbi-analyzer-design.md for the full rule catalog.

pba is built and distributed with uv. If you don't have it yet:

# macOS / Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Homebrew (macOS / Linux)
brew install uv

# Windows (PowerShell)
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Restart your shell (or source ~/.zshrc / ~/.bashrc) and verify:

uv --version

uv tool install git+https://github.com/databricks-solutions/powerbi-analyzer
# or for development
git clone https://github.com/databricks-solutions/powerbi-analyzer.git && cd powerbi-analyzer
uv venv && source .venv/bin/activate
uv pip install -e ".[dev]"
pre-commit install

pba pbix path/to/report.pbix --out reports/audit.md
pba pbix path/to/report.pbip --out reports/audit.html

pba workspace --workspace-id <guid> --tenant-id <guid> --auth device_code
# follow the device-code prompt in your browser

Required Power BI scopes: Dataset.Read.All, Workspace.Read.All, Tenant.Read.All.

For unattended use, set PBI_TENANT_ID, PBI_CLIENT_ID, PBI_CLIENT_SECRET and pass --auth service_principal.

pba databricks --profile DEFAULT --warehouse-id <id> --catalog main.gold

Required permissions:

• USE CATALOG on each scanned catalog
• SELECT on system.query.history, system.compute.warehouse_events, system.information_schema.*
• CAN_USE on the warehouse

pba init                 # writes pba.yaml
$EDITOR pba.yaml         # fill in credentials and targets
pba scan                 # produces a single combined report

Two renderers are supported: markdown (default) and html (single self-contained file with embedded CSS, JS, and DM Sans font — safe to email or drop on a share).

Per-command (pba pbix, pba workspace, pba databricks):

pba pbix report.pbix --out reports/audit.html        # inferred from .html suffix
pba pbix report.pbix --out reports/audit.md --formats html   # or forced via flag

From pba.yaml (used by pba scan):

output:
  formats: [markdown, html]   # any subset; both written side-by-side
  dir: reports/

See [examples/sample-report.md](examples/sample-report.md) and [examples/sample-report.html](examples/sample-report.html) for what the report looks like on the deliberately-bad fixture.

pytest                   # run all tests
mypy src/                # type-check
ruff check src tests     # lint
ruff format src tests    # format
PBA_UPDATE_GOLDENS=1 pytest tests/e2e/  # refresh golden reports

Please attach the contents of ~/.cache/pba/<run_id>/ (collector outputs, redacted) when filing a bug — the cache is sanitized of secrets, GUIDs, and connection strings.

© 2025 Databricks, Inc. All rights reserved. The source in this notebook is provided subject to the Databricks License [https://databricks.com/db-license-source]. All included or referenced third party libraries are subject to the licenses set forth below.

The full Databricks DB License text is in LICENSE.md.

The content provided here is for reference and educational purposes only. It is not officially supported by Databricks under any Service Level Agreements (SLAs). All materials are provided AS IS, without any guarantees or warranties, and are not intended for production use without proper review and testing.

If you encounter issues while using this content, please open a GitHub Issue in this repository. Issues will be reviewed as time permits, but there are no formal SLAs for support.