Synapse to Fabric Case Study: Consolidating Synapse, ADF and Power BI Premium onto Microsoft Fabric
A representative scenario: consolidate in phases, retire after sign-off.
This Synapse to Fabric case study walks through a representative consolidation of Azure Synapse, Azure Data Factory and Power BI Premium onto Microsoft Fabric: why teams do it, how the architecture changes, how to order the work and how to prove the reports still match.
It is a scenario, not a client story. It names no client and quotes no measured client figures, and the technical facts come from Microsoft documentation. Real, measured projects are summarised in the section on our published projects below, and linked to their full case studies.
About This Synapse to Fabric Case Study Scenario
The scenario is a mid-sized organisation with three platforms behind one reporting estate. A Synapse dedicated SQL pool and Spark pools hold the warehouse and transformation code. Azure Data Factory or Synapse pipelines load it. Power BI reports run on a Premium per capacity SKU. Each platform has its own bill, its own admin console and its own skills.
Treat your own estate as the real case. Use this walkthrough as a template and replace each assumption with your own baseline before you plan.
Why Consolidate Synapse, ADF and Power BI Premium
The driver is usually simplicity rather than urgency. Microsoft has announced no retirement for Synapse, but it positions Fabric as the strategic platform for new analytics innovation, and new pipeline features are not backported to Azure Data Factory or Synapse pipelines. Teams that want those features, and one capacity instead of three meters, plan the move.
Our Azure Synapse vs Microsoft Fabric comparison helps with the decision, and data platform consolidation on Microsoft Fabric covers the wider case. If you need the numbers for sign-off, start with the Microsoft Fabric business case.
Synapse to Fabric Case Study Architecture: Before and After
Before the move, data flows through Synapse pools and pipelines into Power BI import models on a Premium capacity. After the move, data sits in OneLake as Delta tables, organised in bronze, silver and gold layers, with Fabric Warehouse or Lakehouse, Spark notebooks, Data Factory pipelines and Power BI all running on one Fabric capacity. Our guide to the medallion architecture in Azure Synapse shows the layer design that carries across.
Representative consolidation, not a specific client.
Each of the three pieces moves differently.
Warehouse and Spark
Fabric's Migration Assistant for Data Warehouse covers Synapse dedicated SQL pools. For Spark, the Spark Migration Assistant is in preview and does not migrate configurations or libraries, and it cannot migrate workspaces that use managed virtual networks, so plan manual work. Calls to mssparkutils change to notebookutils.
Pipelines
Azure Data Factory and Synapse pipelines can be upgraded with the built-in upgrade experience or the Microsoft.FabricPipelineUpgrade PowerShell module, which maps linked services to Fabric connections through a resolution file. It supports a limited set of datasets, so expect some pipelines to be rebuilt. You can also add an Azure Data Factory item to a Fabric workspace for visibility and an incremental upgrade. Our guide to SSIS to Fabric Data Factory covers the same pipeline patterns.
Data in ADLS Gen2
OneLake shortcuts to ADLS Gen2 are the zero-copy route: Fabric reads the existing data in place, so you can move reporting first and copy later, if at all.
Power BI Premium
Premium capacity moves workspace by workspace, with no tenant-wide switch. Microsoft's sequence is to inventory, buy the F SKU first, pilot a low-impact workspace, reassign workspaces, validate and then cancel the P SKU. Same-region reassignment is the default, and Fabric items do not survive cross-region moves.
Approach and Timeline
The plan below shows the order of work, not durations. Set durations from your own inventory, because the number of pipelines, models and reports drives the schedule far more than data volume.
Illustrative order of work.
| Phase | What happens | Exit signal |
|---|---|---|
| 1. Assess | Inventory pools, pipelines, models, reports and costs; pick a first domain | Scope and baseline recorded |
| 2. Land | Create OneLake shortcuts to ADLS Gen2; set workspaces and security | Data readable in Fabric |
| 3. Move | Migrate warehouse, Spark and pipelines; upgrade pipelines where possible | Workloads run on Fabric |
| 4. Validate | Parallel run through a full reporting cycle | Written parity sign-off |
| 5. Cut over | Reassign Power BI workspaces to the F SKU; monitor capacity | Users on Fabric |
| 6. Retire | Pause Synapse pools and cancel the P SKU once sign-off is given | Cost and scope closed |
Move by business domain, not by technology layer. Our Microsoft Fabric migration checklist keeps each wave consistent, and the guide to migrating to Microsoft Fabric from any platform covers other sources. For a similar scenario from a different source, see our Snowflake to Fabric case study.
Sizing the Fabric Capacity and Licences
Premium features such as viewing without a Pro licence are available only on F64 and larger. Smaller F SKUs need Pro licences for viewers, so the SKU decision changes the licence bill as well as the capacity bill. Our Microsoft Fabric SKU guide for executives explains the trade-offs.
Pay-as-you-go is typically higher than the equivalent P SKU, so evaluate reservations and pause and resume for Synapse-style workloads. After reassigning workspaces, watch the Capacity Metrics app for 24 to 48 hours. Premium per capacity SKUs are being retired on a published timeline, so check your agreement dates early and plan overlap rather than a last-day switch.
Parity Validation: How to Prove the Numbers Match
Reports are trusted because the numbers agree with what people already know. Validate in three layers, and keep the evidence.
Layer 1: Tables
Compare row counts and checksums of key columns for each migrated table, and check null counts and date ranges.
Layer 2: Measures
Compare the business measures that drive decisions by period and by key dimension. Agree the tolerance with the business owner before testing, so a rounding difference is not argued afterwards.
Layer 3: Reports
Open the old and new report side by side for the same filters, including a month-end close, and record any difference with its cause. Test security by signing in as users from different roles. For a Premium move, run old and new in parallel for a short window before reassigning the remaining workspaces.
- Consolidate for simplicity and features; Synapse has no announced retirement.
- Use ADLS Gen2 shortcuts so data is read in place before anything is copied.
- Upgrade pipelines with the tooling, and plan to rebuild the unsupported ones.
- Move Premium workspace by workspace: buy the F SKU first, cancel the P SKU last.
- Prove parity at table, measure and report level before you retire anything.
Synapse to Fabric Case Study Results: What to Measure
A real case study reports measured results. This scenario deliberately does not, because any figure here would be invented. Set these baselines before the migration and measure the same things afterwards.
| Metric | How to measure it | Evidence |
|---|---|---|
| Platform cost | Synapse, Data Factory and Premium spend before, against Fabric capacity and storage after | Invoices for the same months and scope |
| Report load time | Median and slowest-visual time for the same reports | Performance Analyzer on both platforms |
| Refresh duration | Time for the same semantic models to refresh, or move to Direct Lake | Refresh history |
| Parity | Share of checked measures inside tolerance | Parity test log |
| Number of platforms | Billing meters and admin consoles in use | Platform inventory |
| Incidents | Failed loads and report errors per month | Monitoring and ticket history |
Compare like with like: the same months, the same reports and the same scope. Include costs you keep, because a partial move can leave old and new bills in place for a while.
What Our Published Projects Measured
For real figures, see the client projects on our case studies page. The numbers below are self-reported by those projects and the clients are anonymised.
Power BI Premium to Fabric, UK investment manager
This project moved 214 reports from a Premium P3 capacity to Fabric F64 and 15 Azure Data Factory pipelines to Fabric Data Factory. It reports a 55% licence cost reduction and refresh falling from 45 minutes to under 90 seconds after a move from import to Direct Lake. Read the full Power BI Premium to Fabric case study.
NHS integrated care board
This project migrated 17 of 22 Azure Data Factory pipelines and rebuilt 5 on a Fabric F32 capacity. It reports data latency falling from 8 hours to 20 minutes and 62% fewer ad-hoc requests. Read the full NHS integrated care board case study.
Lessons Learned from Synapse to Fabric Migrations
Do not assume a lift and shift
Some pipelines, Spark configurations and libraries need manual work. Find them in the inventory, not during cutover.
Keep regions aligned
Reassigning a workspace to a capacity in another region is not a simple switch, and Fabric items do not survive cross-region moves. Check regions before you buy the F SKU.
Do not retire too soon
Keep the old platforms available for a rollback window after cutover, and check contract and renewal dates early so the timeline and the dates align.
Keep what still fits
Not everything has to move. Microsoft states it has no plans to deprecate Azure Analysis Services, so a stable model there can stay while you move the rest.
Synapse to Fabric Case Study FAQ
Is this a real client case study?
No. It is a representative scenario built from how Synapse, Data Factory and Power BI Premium consolidations are usually structured, using Microsoft documentation for the facts. It names no client and quotes no measured client figures. Real projects are on our case studies page.
Is Azure Synapse being retired?
Microsoft has not announced a retirement date for Azure Synapse Analytics. It describes Fabric as its strategic platform for new analytics innovation, so many teams plan consolidation on their own timeline rather than a deadline. Check the current Microsoft guidance before you set dates.
Can Data Factory pipelines be upgraded to Fabric automatically?
Partly. Azure Data Factory and Synapse pipelines can be upgraded with the built-in upgrade experience in the Data Factory UI or with the Microsoft.FabricPipelineUpgrade PowerShell module, which uses a resolution file to map linked services to Fabric connections. Only a limited set of datasets is supported, so some pipelines need rebuilding.
How do Power BI Premium workspaces move to Fabric?
Workspace by workspace. You buy the Fabric F SKU first, pilot a low-impact workspace, reassign workspaces from the P SKU capacity, validate, and only then cancel the P SKU. Premium features are available on F64 and larger, and smaller F SKUs need Pro licences for viewers.
Does the Spark Migration Assistant move everything?
No. The Spark Migration Assistant is in preview. It does not migrate Spark configurations or libraries and cannot migrate workspaces that use managed virtual networks. Code that uses mssparkutils needs to change to notebookutils.
How do I prove the numbers match after the move?
Run both platforms in parallel for at least one full reporting cycle, such as a month-end close. Compare row counts, key aggregates and the finished reports, agree a tolerance with the business owner first, and keep the evidence for sign-off.
To plan your own move, see our guide to choosing a Microsoft Fabric migration partner, our Azure Synapse consulting and our Microsoft Fabric migration services, or request a free migration assessment.