Skip to main content

Align assets to prioritized workloads in Microsoft Fabric

In Microsoft Fabric, a workload is a conceptual unit of data and analytics processing composed of artifacts like Lakehouses, Warehouses, Power BI datasets, Dataflows, Pipelines, Eventstreams, and Real-Time Hubs. The previous article, Define and prioritize workloads, provided guidance for identifying key workloads. This article supports the next step: aligning Fabric assets and resources to those prioritized workloads before implementing them.

The following technical inputs should be validated and extended for Fabric-based projects:

  • Fabric artifacts: List Lakehouses, Warehouses, KQL Databases, Power BI artifacts, Pipelines, Eventstreams, and Notebooks included in the workload.
  • Capacities: Specify the Fabric capacities the workload consumes or requires, including SKUs and estimated utilization.
  • Data sources and destinations: Identify internal and external data sources accessed (e.g., SQL DBs, APIs, OneLake Shortcuts), and destination storage (e.g., OneLake, Power BI, external sinks).
  • Dependencies: List upstream/downstream datasets, pipelines, triggers, and inter-service dependencies (e.g., with Azure Data Factory, Synapse, or Power Platform).
  • Access and governance: Identify security groups, sensitivity labels, workspace structure, and lineage requirements.

Asset Alignment by Adoption Type​

Your prioritization effort should consider the type of adoption activity:

Migrate​

In migration-focused efforts, you bring existing analytical workloads into Fabric with minimal rework. For example:

  • Power BI reports moving from legacy shared datasets into centrally governed Fabric workspaces.
  • SQL-based ETL processes rehosted using Dataflows Gen2 or T-SQL Pipelines.
  • Azure Data Explorer (ADX) solutions shifting to KQL databases in Fabric.

Effort estimation focuses on the number and size of artifacts, complexity of refresh schedules, and compatibility with Fabric capacities.

Modernize​

Modernization introduces managed components of Fabric (e.g., Lakehouses replacing SQL DWs or data marts). In this context:

  • Add indicators of refactoring effort (e.g., converting SSIS packages to Data Pipelines).
  • Assess opportunities to consolidate scattered artifacts across workspaces.
  • Track optimization goals like data volume reduction, semantic model reuse, and automation via Notebooks.

Innovate​

For workloads centered on innovation (e.g., integrating Real-Time Intelligence, AI Copilot, or cross-domain data products):

  • Focus less on migrated artifacts and more on new architectural patterns.
  • Track inter-service orchestration (e.g., Eventstreams feeding KQL DBs and pushing outputs to Power BI Dashboards).
  • Emphasize governance challenges, cost scaling, and data contracts.

Automation and Tooling​

Fabric currently does not support automated grouping like Azure Migrate, but similar efficiencies can be achieved through:

  • Microsoft Purview: To track lineage, sensitivity, and usage of Power BI and Fabric artifacts.
  • Workspace and Capacity APIs: For scripting inventory across workspaces.
  • Fabric Monitoring Hub: To assess activity per workspace, job failures, and execution time for pipelines or notebooks.

Configuration Management Integration​

If your organization maintains a Configuration Management Database (CMDB) or enterprise catalog (e.g., ServiceNow, Collibra, Microsoft Purview), integrate workload alignment by:

  • Tagging Fabric assets with CMDB IDs or project codes.
  • Mapping business units or owners directly within Fabric workspaces.
  • Embedding classification or risk level metadata into Lakehouse folders or Power BI datasets.

Once assets are aligned, your Fabric adoption plan becomes ready for iterative workload delivery cycles and FinOps planning.

Contributors