Skip to main content

Evaluate Fabric Workload Readiness

When preparing to migrate data workloads into Microsoft Fabric, it's essential to evaluate the readiness of each workload. This includes checking for compatibility with Fabric’s architecture, data formats, processing patterns, and governance policies.

Evaluation assumptions​

While the Cloud Adoption Framework often provides cloud-agnostic principles, readiness evaluations for Microsoft Fabric are specific. Assessments should include:

  • Data format alignment (e.g. CSV, Parquet, Delta Lake compatibility)
  • Ingestion readiness via pipelines or Eventstream
  • Compatibility with Fabric Lakehouses, Warehouses, or Real-Time Hubs
  • Identity integration via Entra ID (Azure AD)
  • Cross-service dependencies with Power BI, Data Activator, and Synapse (for legacy scenarios)

Evaluate cross-datacenter dependencies​

If you're migrating from distributed data landscapes or hybrid models:

  • Use tools like Microsoft Purview for lineage and dependency mapping.
  • Use Azure Migrate to identify upstream and downstream application dependencies.
  • Assess bidirectional dependencies to avoid broken chains during staged migrations.

Evaluate readiness for Microsoft Fabric SQL Workloads​

For SQL-based workloads, especially those targeted for SQL Endpoint in Fabric Lakehouse, evaluate these:

  • Query compatibility: T-SQL dialect in Fabric supports most standard operations. See T-SQL differences in Fabric
  • Performance expectations: Use Workload Analyzer to simulate peak usage.
  • Schema complexity: Normalize where necessary and assess for Fabric best practices like wide tables and star schema designs.
  • Availability & RPO/RTO: Document expectations and align with Fabric's zone redundancy and OneLake storage durability.
  • Size and update patterns: Identify large databases or those with high change rates. These require specific pipeline strategies or snapshot logic.

📘 For a Fabric-centric checklist on database compatibility and migration, refer to SQL migration readiness

Network and Bandwidth Considerations​

  • Estimate OneLake write and read throughput
  • Consider Data Gateway or Hybrid Connections if legacy on-premises components remain
  • Include bandwidth requirements for parallel pipeline executions, event ingestion, and analytics queries

Bandwidth Estimation Techniques​

  • Initial sync size: Total volume of data to ingest into Fabric
  • Ongoing delta rate: Expected growth rate of each workload
  • Expected concurrency: Number of parallel ingestion paths and data flows
  • Use Azure Speed Test for initial estimates between your datacenter and Fabric region

Final Notes​

Align the readiness evaluation with your Fabric Landing Zone standards, and make sure all workload assessments feed into your Fabric backlog tracking system (e.g. Azure DevOps).

For deeper assessments, see:

Contributors