Outcome: a single Pull Request, opened by your team, exercises every challenge you completed — workspace, items, domain, label, access, medallion, agent, audit, promotion — and you can prove it all worked from the dashboard.
Goal
Pick a realistic business scenario and ship it through your governance plane in one coordinated PR (or a short series of stacked PRs). Then demo the artifacts to the judges:
- Manifests committed.
- Validate workflow green.
- Provision workflow green (with the
productionenvironment gate visibly approved). - Live Fabric resources matching the manifests.
- Governance dashboard showing the new activity.
- Drift detector showing zero drift.
- (Bonus) An Activator alert that fired on something interesting.
Suggested scenario
"Finance Revenue Reporting goes live in production"
A new business line in Finance needs to ship a certified revenue dashboard across dev → stg → prd in two weeks. They need ingestion from Azure SQL into a bronze lakehouse, dedup + standardize in silver, build a certified gold semantic model, and expose a Fabric Data Agent for executives.
Adapt freely. The scenario just exists to keep your team's PR coherent.
Required artifacts
Map each row to the relevant challenge(s).
| # | Artifact | Challenge(s) |
|---|---|---|
| 1 | workspaces/dev-fin-revenue.yaml, workspaces/stg-fin-revenue.yaml, workspaces/prd-fin-revenue.yaml |
01 |
| 2 | domains/finance.yaml, capacities/<your-capacity>.yaml |
03 |
| 3 | items/<workspace>/lh_bronze_*.yaml, lh_silver_*.yaml, lh_gold_*.yaml |
02 |
| 4 | items/prd-fin-revenue/sm_gold_revenue.yaml (Certified) |
02, 05 |
| 5 | access/<workspace>.yaml for each workspace, group-first, expiring user bindings |
04 |
| 6 | medallion/finance-revenue.yaml |
05 |
| 7 | agents/agent-finance-revenue-qa.yaml + instructions file |
06 |
| 8 | Eventhouse + Eventstream + dashboard already provisioned in dev-plt-gov-audit |
07 |
| 9 | pipelines/finance-revenue.yaml + a promote/stg PR and a promote/prd PR |
08 |
You don't have to do every row — finishing any 5 is a strong submission.
Judging rubric
100 points total.
| Area | Points | What we look for |
|---|---|---|
| Correctness | 25 | Manifests pass validate; resources actually exist; labels and endorsement applied. |
| Coverage | 20 | Number of challenge areas integrated into the PR chain. |
| Governance discipline | 20 | No out-of-band portal clicks; managed-by: marker present everywhere; drift = 0. |
| MCP + Skills usage | 15 | Evidence the team used both Fabric MCP and at least 2 Skills in the workflow. |
| Observability | 10 | Dashboard reflects this scenario's activity; at least one Activator alert wired. |
| Story | 10 | Demo articulates the business benefit, not just the plumbing. |
Demo expectations
A ~10-minute walkthrough that includes:
- The original PR (open in GitHub).
- The sticky validate comment (passing).
- The provision run (with environment approval timestamp).
- The Fabric portal (showing the workspaces and items).
- The governance dashboard (one filter on the team's pipeline).
- The agent answering a sample question.
- Optional: an injected drift event, with the detector catching it.
Stretch (if you have spare time)
- Demonstrate the rollback flow on prd (Challenge 08).
- Demonstrate the break-glass access path (Challenge 04) with the automatic expiry follow-up PR.
- Show a Purview DSPM signal blocking an agent (Challenge 06).
- Add the team's own new policy rule (e.g., "no Lakehouse may be unowned by
a Group") to
rules/policy.yamlwith tests.
References
Every previous challenge — keep them open in tabs during the capstone.