Deployment Pipelines: Dev → Test → Prod

What are deployment pipelines?

Simple explanation

Think of deployment pipelines like a quality control line in a factory.

A new product (report, model) starts in the development workshop. After testing, it moves to the test floor for quality checks. Once approved, it goes to the production warehouse where customers (business users) access it.

Fabric deployment pipelines automate this: Dev workspace → Test workspace → Production workspace. Each promotion copies items and can swap data connections (dev data → prod data).

Pipeline stages

Stage	Purpose	Audience
Development	Build and iterate on new features	Data engineers, report developers
Test	Validate with production-like data	QA team, business stakeholders
Production	Live environment for business users	All report consumers

Creating a deployment pipeline

Go to Fabric portal → Deployment pipelines
Create pipeline and name it
Assign workspaces to each stage (Dev, Test, Prod)
Compare — the pipeline shows differences between stages
Deploy — promote items from one stage to the next

What gets deployed

All Fabric items can be deployed: semantic models, reports, dashboards, lakehouses (metadata), warehouses (metadata), dataflows, pipelines, notebooks.

Deployment rules

Deployment rules change item parameters per stage — crucial for data source separation:

Rule Type	What It Changes	Example
Data source	Connection string or server name	Dev connects to `sql-dev.fabric.com`, Prod connects to `sql-prod.fabric.com`
Parameter	Power Query M parameter values	`ServerName = "dev-server"` → `"prod-server"`
Lakehouse	Which lakehouse an item points to	Dev lakehouse → Prod lakehouse

Scenario: James promotes a client report

James at Summit Consulting builds a new revenue dashboard in the Dev workspace using test data. After review:

He deploys from Dev → Test (the pipeline copies the report and model)
Deployment rules swap the data connection from lakehouse-dev to lakehouse-test
The QA team validates with production-like data
James deploys from Test → Prod (data connection swaps to lakehouse-prod)

Business users see the new dashboard in the Prod workspace — connected to real data — without any manual configuration.

Exam tip: Deployment pipelines vs Azure DevOps

The exam may ask you to differentiate:

Fabric deployment pipelines = Fabric-native, workspace-to-workspace promotion, UI-based
Azure DevOps / GitHub Actions = CI/CD automation, code-based, can call Fabric REST APIs
Git integration = version control, tracks changes, enables branching

These tools complement each other: Git tracks changes, deployment pipelines promote items, and CI/CD automates the process.

Question

What do Fabric deployment pipelines do?

Click or press Enter to reveal answer

Answer

Promote items across workspace stages (Dev → Test → Prod). They compare stages, show changes, and deploy selected items. Deployment rules swap data connections per stage.

Click to flip back

Question

What are deployment rules?

Click or press Enter to reveal answer

Answer

Rules that change item parameters per stage. Example: data source connection swaps from dev-server to prod-server during deployment. This ensures each stage uses appropriate data without manual reconfiguration.

Click to flip back

Knowledge Check

James deploys a semantic model from Dev to Prod. The model connects to lakehouse-dev. What ensures it connects to lakehouse-prod after deployment?

Question

What can deployment rules change per stage?

Click or press Enter to reveal answer

Answer

Deployment rules change three types of parameters when items move between stages: (1) Data source connection strings — swap dev to prod server, (2) Power Query M parameter values — change server names or file paths, (3) Lakehouse references — point items to a different lakehouse per stage. Rules are configured once and applied automatically on every deployment.

Click to flip back

Knowledge Check

James deploys a report from Dev to Test. The report's semantic model uses lakehouse-dev. In Test, the data should come from lakehouse-test. What enables this automatic swap?

Next up: Impact Analysis & Dependencies