Skip to content
Projects

microsoft/fabric-architecture-review

View on GitHub

microsoft/fabric-architecture-review has 28 GitHub stars and 12 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-12 07:28 UTC. Last push 2026-11-05.

How these counts are measured

An accelerator that reviews a Microsoft Fabric tenant end-to-end through metadata-only scans across governance, cost, capacity, performance, and security, delivered as a consultant report, Power BI governance model, conversational data agent, estate knowledge graph, and standalone Rayfin-powered interactive workbench.

What you can do

  • FUAM and FCA: enrich the review with activity, capacity and sharing evidence. Without these connections, their rules are not applicable; FAR does not collect replacement monitoring or financial data.
  • Workspace-owner reporting: provide a separate, secured view for approved administrators. Follow the owner access guide.
  • Targeted reviews: use FUAM to select workspaces and optionally notify their administrators. See targeted reviews.
  • Agents, app and ontology: explore results conversationally or visually. See Fabric results and app setup.

Measured star growth

in 24h
0
in 7 days
Collecting history
in 15 days
Collecting history
in 1 month
Collecting history

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

README.md

Original

Fabric Architecture Review (FAR)

License: MIT Python 3.11+

Assess the architecture, governance and security of your Microsoft Fabric estate. FAR reviews workspace design, configuration and access, then explains each finding and the next action to take. It does not query customer business rows.

Explore results in Power BI, the FAR app, a Data Agent or local Markdown/PDF reports. Optional workspace-owner reports restrict each reader to their authorized workspaces.

Get started

Run in Start here
Microsoft Fabric Deploy FAR: import setup, configure the review and run the pipeline
Your workstation Run a local review: install dependencies, create .env.far and run the three stages

Start with one test workspace. An empty workspace scope includes all workspaces the review identity can access.

See the 2026.11.0 release notes for this version.

What FAR reviews

70 rules cover five dimensions. Only assessed controls affect the score. Signals highlight areas to investigate; missing evidence is never a pass.

Dimension Examples
Architecture Workspace boundaries, hosting, storage modes, pipeline/notebook contracts and dependencies
Governance Administrator continuity, naming, documentation, endorsement and sensitivity labels
Operations Source control, deployment pipelines, capacity state and optional activity signals
Reliability Production concentration, pipeline failure paths and dependencies on non-production
Security Sharing, direct access, tenant settings, network controls, OneLake roles and semantic-model RLS/OLS

Declare production workspaces and required policies before interpreting results. FAR does not infer your requirements from workspace names. See the rule catalogue and scoring guide.

Configure the review

Setting Microsoft Fabric Local
Review identity FAR_Environment Variable Library .env.far
Optional FUAM/FCA connections FAR_Connections Variable Library sources.yaml
Thresholds and policy FAR_Thresholds Variable Library thresholds.yaml, policy.yaml
Workspace scope and report labels Pipeline parameters .env.far

Setup preserves configured values. Follow the FAR policy guide for examples, semantic-model security checks and optional role details.

Optional capabilities

  • FUAM and FCA: enrich the review with activity, capacity and sharing evidence. Without these connections, their rules are not applicable; FAR does not collect replacement monitoring or financial data.
  • Workspace-owner reporting: provide a separate, secured view for approved administrators. Follow the owner access guide.
  • Targeted reviews: use FUAM to select workspaces and optionally notify their administrators. See targeted reviews.
  • Agents, app and ontology: explore results conversationally or visually. See Fabric results and app setup.

Use the results safely

Start with the report's Action plan. Resolve evidence gaps, review failures, and use Items to review to identify affected items. Re-run the same scope after making changes. Compare results only when the review version and configuration match.

The central report, app and Agent are for the review team, not an owner-only audience. Review metadata can be sensitive even though business rows, source code and filter expressions are not collected. Read the permissions guide and data-safety guidance before deployment or sharing.

Validate access with actual users before production use. For assistance, see support.

Explore related projects