Fabric Gold Lakehouse
Fabric Gold Lakehouse
Overview
The Gold layer represents the final, trusted zone in the data pipeline architecture. It hosts business-ready, aggregated, and validated datasets optimised for reporting, analytics, and downstream consumption via Power BI, semantic models, and external applications.
Gold Lakehouse summary
| 📋 Data | Tables and Views stored in schemas. Data is cleaned, transformed, and ready for reporting, analytics, and consumption. |
|---|---|
| 🔄 Transformation | Data is processed and moved using Python and Spark notebooks. |
| 🔐 Access | Database roles enforce access. No access for external applications. User groups can read specific tables and views via the SQL endpoint. |
Organisation, access and governance
All data is stored in a single Lakehouse — spl_gold_lh — which is organised using schema separation, with schema names aligned to source domains for clarity and governance.
To support organisation-wide sharing and simplify access management, shortcuts are no longer recommended in this Lakehouse. Avoiding shortcuts eliminates the need for additional permissions on the underlying Lakehouses, reducing complexity and potential access issues.
Access to spl_gold_lh is tightly controlled:
- Only designated user groups — such as PBI Power Users — are granted SQL-endpoint access.
- Schema-level and object-level permissions are enforced using database roles, ensuring fine-grained control over data exposure and operational integrity.
Architecture
flowchart LR
SIL[(spl_silver_lh
Silver — transformed)] -->|Python / Spark notebooks| GOLD[(spl_gold_lh
Gold — trusted)]
GOLD --> SQLE[SQL endpoint]
SQLE -->|database roles| RG1[PBI Power Users]
SQLE -->|database roles| RG2[Other designated groups]
GOLD --> PBI[Power BI
reports and semantic models]
GOLD --> EXT[External applications
read-only via SQL endpoint]
classDef sil fill:#c0c0c0,stroke:#555,stroke-width:1px,color:#000
classDef gold fill:#ffd54f,stroke:#b45f06,stroke-width:1px,color:#000
classDef sec fill:#ffe0b2,stroke:#b45f06,stroke-width:1px,color:#000
classDef cons fill:#e3f2fd,stroke:#1976d2,stroke-width:1px
class SIL sil
class GOLD,SQLE gold
class RG1,RG2 sec
class PBI,EXT cons
Comparison Bronze / Silver / Gold
| Bronze | Silver | Gold | |
|---|---|---|---|
| Purpose | Raw landing zone | Transformation zone | Trusted, business-ready |
| Lakehouses | 5x (BMD, Rimo, Board, SharePoint, H3A) | 1x spl_silver_lh | 1x spl_gold_lh |
| Shortcuts allowed | n/a — sources | Yes — referencing Bronze | No — sharing/governance reason |
| Transformations | None | Cleansing, joining, enrichment | Aggregations, business-ready datasets |
| Consumers | Fabric team | Fabric team / pipelines | Power BI, semantic models, external apps (read-only) |
| Access control | Restricted to Fabric team | Workspace members | Database roles (group-based) |
Schemas and objects in spl_gold_lh
Views in spl_gold_lh
Shows the view definition and dependencies.
Report: Views in spl_gold_lh (Power BI)
Operational notes
- No shortcuts: Gold vermeidet bewusst Shortcuts. Tabellen werden persistent geschrieben, um Sharing und Governance zu vereinfachen.
- SQL endpoint als zentraler Zugang: Externe Anwendungen und Power Users lesen ueber den SQL endpoint, nicht ueber die Lakehouse-API.
- Database roles: Berechtigungen werden auf Schema- und Objekt-Ebene per Rolle vergeben — nicht pro User.
- Semantic models: Power BI semantic models konsumieren aus Gold und stellen die letzte Modellierungs-Schicht fuer Reports dar.
Related pages
- Bronze Lakehouse (Hub)
- Fabric Silver Lakehouse
- Fabric Lakehouse Hub
- Microsoft Fabric Datenarchitektur im SPL
Verwandte Themen
Kuratiert — kann direkt im Editor ergänzt werden://
Automatisch vorgeschlagen über gemeinsame Tags (architektur, fabric):
- BMD Bronze Lakehouse
- Board Bronze Lakehouse
- H3A Bronze Lakehouse
- Rimo Lakehouse
- SharePoint Bronze Lakehouse
- Bronze Lakehouse
Status: Migriert — Team-Review ausstehend · Owner: (festlegen) · Letzter Review: 2026-06-11