Fabric Data Platform
An end-to-end Azure and Microsoft Fabric analytics blueprint: Bicep infrastructure, streaming ingestion, SQL, KQL and M transformations, ML and governance, with tested offline samples.
- 66Components mapped
- 3Clouds as sources
- Bronze → GoldMedallion layers
Overview
An end-to-end analytics repository blueprint for Azure and Microsoft Fabric, mapping every part of a reference architecture to working code, templates and configuration.
It covers multicloud sources, ingestion, a OneLake lakehouse, warehouses, real-time intelligence, semantic models, AI and governance. All 66 diagram components are mapped to repository artifacts, with offline samples and tests that run without an Azure account.
This repository is a blueprint with working offline samples. It has not been deployed to a live tenant.
What it does
Infrastructure as code
Bicep for Fabric capacity, storage, streaming, AI/ML, Purview, Key Vault, networking, policy and budget.
Multicloud sources
CloudFormation for S3 and Kinesis on AWS, Terraform for Cloud Storage and Pub/Sub on Google Cloud.
Lakehouse notebooks
Bronze, Silver and Gold processing plus machine learning source code.
Real-time intelligence
Eventstream and Eventhouse definitions with KQL schema, mappings and queries.
Semantic model
A Direct Lake model with DAX measures for Power BI reports.
Governance & monitoring
Access matrix, Purview operating model and alert contracts.
How it works
Ingest
Mirroring, Data Factory, Dataflow Gen2, shortcuts and Eventstream bring data into OneLake.
Store
Lakehouse, warehouse, SQL database and Eventhouse hold the data.
Process
Notebooks, stored procedures and dataflows transform it from Bronze to Gold.
Serve
Power BI, real-time dashboards, SQL endpoints, GraphQL and AI agents consume it.