Forum Discussion
Navigating through Forums and Migrating from Synapse to Azure Fabric
- 10 months ago
Hi XhevahirMehalla,
Thanks for reaching out to the Microsoft Fabric Community Forum.
We understand you’re looking for guidance on both navigating Microsoft community forums and assessing your migration from Azure Synapse Analytics to Microsoft Fabric. Let’s address each of your queries in detail below.You can post your questions in the following Microsoft community forums:
Power BI forums - Microsoft Fabric Community
Fabric platform forums - Microsoft Fabric Community
Microsoft Fabric brings together data engineering, data movement, data science, real-time analytics, and business intelligence within a single SaaS platform. Transitioning from Synapse means moving your pipelines, datasets, and workloads to Fabric’s unified environment, such as Lakehouse or Warehouse. Since your current setup mainly uses Copy Data, Data Flows, and Azure SQL Database and not Spark or SQL Serverless, the migration should be of moderate complexity.
Data Ingestion:
In Fabric, you use Data Pipelines or Dataflow Gen2 for ingestion, both offering built-in connectivity to Oracle and ADLS Gen2. The Fabric Gateway now replaces the Integration Runtime (IR) for connecting to on-premises or private sources, so you'll need to set up your Oracle connections and gateway in your new Fabric workspace.Data Transformation:
You can rebuild your Synapse Data Flows with Dataflow Gen2 in Fabric, using similar visual mapping and transformation logic. Alternatively, you can use Notebooks (Spark) for more advanced or scalable data transformations.Data Storage:
Fabric centralizes data in OneLake, a unified data lake. You can link your existing ADLS Gen2 containers as Shortcuts to avoid data duplication and set up your raw and staging zones in a Fabric Lakehouse with minimal changes.Data Loading:
If you load data into Azure SQL DB, you can move this to Fabric Warehouse, which has a managed T-SQL endpoint, supports incremental loads, and integrates with Power BI, providing a familiar relational experience and unified Fabric features.Application Integration:
Applications that upload Excel or CSV files can still do so directly into OneLake using REST APIs or Shortcuts. Automated downstream processes can continue with Fabric Pipelines or Dataflows, requiring little to no logic changes.DevOps and CI/CD:
Fabric offers built-in Git integration with Azure DevOps and GitHub, allowing version control for pipelines, lakehouses, and reports. You can use Fabric Deployment Pipelines to promote workspaces, streamlining your Synapse release process with a SaaS-based approach.VNETs and Private Endpoints: Fabric is a SaaS service, but private access is supported using Private Links (in preview) and Managed VNETs for secure connectivity.
DNS and Private Zones: These can continue as-is; you only need to ensure Fabric workspaces have access via Fabric Gateway.
Multi-Subscription Setup: You can replicate your Dev/UAT and Prod setup with separate Fabric workspaces and managed capacities.
Pros:
Unified data platform (ETL, analytics, BI in one environment)
Native Power BI integration and data governance through OneLake
Reduced infrastructure maintenance (no IR or manual scaling)
Built-in DevOps, semantic modeling, and monitoring features
Simplified licensing with pay-per-capacity model
Challenges:
Requires rebuilding Synapse Data Flows and Pipelines in Fabric equivalents
Limited support for some complex Synapse activities (e.g., stored procedures, parameterized pipelines)
Learning curve for Fabric concepts (Lakehouse, Warehouse, Dataflow Gen2)
Fabric is still evolving, so some advanced enterprise network controls may be in preview
Since your current architecture doesn’t involve Spark, Serverless SQL, or complex ETL logic, the migration is medium-effort.
Time Estimate: 3–5 weeks for end-to-end reimplementation and testing.
Resources Needed: 1–2 Data Engineers, 1 Architect, and 1 QA for validation.
Skill Readiness: Your Synapse team can easily adapt since Fabric uses similar data pipeline concepts and SQL-based transformations.
If you prefer to stay on Synapse:
You can copy the existing code to another tenant using ARM templates or Azure DevOps release pipelines. This approach minimizes rework but won’t leverage Fabric’s unified features or future roadmap. For clients requiring modernization and cost optimization, Fabric offers a more sustainable direction.
Given your background with Fabric and your position as the architect, the best way forward is to start with a pilot migration by moving one of your simpler Synapse pipelines completely into Microsoft Fabric. This lets you test for performance, cost, and governance within your client’s setup before a full migration. If the first phase meets your expectations, you can continue migrating in stages, keeping Synapse available as a backup during the process. Once you’ve confirmed stability and user adoption in Fabric, you can retire the Synapse workloads and use Fabric as your main data platform going forward.
Also Thank you lbendlin and GeraldGEmerick for your response.
Regards
CST Member.
Hi XhevahirMehalla,
Thank you for your followup
Your concern is understandable, and starting with Synapse Pipelines is a good strategy. Currently, there aren’t any direct migration scripts for a “lift-and-shift” approach, so the recommended method is to rebuild your Synapse workloads natively in Fabric using similar components. Start with a pilot pipeline, such as a simple Copy Data process, and recreate it in a Fabric Data Pipeline. The interface is similar to Synapse, so activities like Copy, Lookup, ForEach, and Conditional Split should be familiar. Instead of Integration Runtime, you’ll use the Fabric Data Gateway to securely connect to your Oracle OCI source. This allows you to map your Oracle and ADLS Gen2 datasets in Fabric without changing your data flow logic.
For transformations, rebuild your Synapse Mapping Data Flows with Dataflow Gen2 in Fabric, which offers the same visual interface and functions like joins, filters, derived columns, and schema drift, all within Fabric. You can use Shortcuts in OneLake to reference your existing ADLS Gen2 storage, saving time and avoiding duplication. After verifying your raw and staging data, connect the output to a Fabric Lakehouse or Warehouse to keep your SQL-based loading logic consistent.
This step-by-step migration lets your team validate each part while keeping Synapse as a fallback. Once the pilot pipeline works as expected, you can move other workloads, use Fabric’s Git integration with Azure DevOps for version control, and gradually retire Synapse pipelines. Since your setup mainly uses Copy and Transformation activities and not much Spark, the migration should be manageable for your team.
Thank you and continue using Microsoft Fabric Community Forum.
- v-sgandrathi10 months ago
Community Support
Hi XhevahirMehalla,
As we did not get a response, may I know if the above reply could clarify your issue, or could you please help confirm if we may help you with anything else?
Thank you.