Edit

Share via

Fabric Graph overview (preview)

Note

This feature is currently in public preview. This preview is provided without a service-level agreement, and isn't recommended for production workloads. Certain features might not be supported or might have constrained capabilities. For more information, see Supplemental Terms of Use for Microsoft Azure Previews.

Fabric Graph helps you model, visualize, and analyze complex relationships within your data. It's a scalable, enterprise-grade solution that turns disconnected data into AI-powered insights. By using Graph, you can uncover hidden connections within your data and enhance decision-making capabilities.

Unlike traditional relational databases that often require costly joins and complex queries, Graph:

Together, these capabilities enable advanced graph analytics directly on OneLake without the need to manually set up fragile ETL (extract, transform, load) or data replication workflows that break easily when data changes.

Graph automatically scales to handle large workloads, so you can analyze billions of relationships without slowdowns. You can add descriptive tags and details to both items (nodes) and their connections (edges), making it easier to organize and search complex relationships.

By using native GQL and Natural Language to GQL (NL2GQL) support, you gain standards-based querying capabilities optimized for graph operations. These capabilities deliver portability and consistency across graph solutions, so you can migrate queries from other GQL-compliant systems. Graph removes the complexity of joins and transformations to unlock seamless graph analysis and advanced insights at scale, all while using your existing data in OneLake.

Why graph analytics matter

Traditional relational and tabular data formats make it difficult - if not impossible - to map relationships between different data points. For example, these formats can't show the intertwined connections between users, posts, comments, forums, and tags on a social media platform. Graph enables you to uncover hidden connections, communities, and influence within your data. By using Graph, you can answer complex questions about social networks, business processes, and more.

Graph provides an efficient way to model, visualize, and query these relationships. It helps you understand the interconnectedness of your data and drive better insights.

  • Business user: Visually explore relationships, run NL (natural language) queries, and gain insights effortlessly.
  • Data engineer: Define graph models, unify data in OneLake with low and no-code tools.
  • Data scientist: Use graph algorithms and ML (machine learning) in Fabric's data science environment.
  • Developer: Build AI agents and real-time apps using graph-powered contextual insights.

Graph broadens access to graph insights beyond specialized roles. Any user can use connected data in daily decision-making.

What you can do with Graph

By using Graph, you can:

  • Create a labeled property graph over structured data in OneLake by defining its nodes and edges in terms of underlying tabular data.

    Important

    Graph currently doesn't support schema evolution. After you ingest and model your data, the structure of nodes, relationships, and properties is fixed. If you need to make structural changes - such as adding new properties, modifying labels, or changing relationship types - you must reingest the updated source data into a new model.

  • Query by using GQL (Graph Query Language), including pattern matching, path constructs, aggregations, and other features as they're released. The official International Standard for GQL is ISO/IEC 39075 Information Technology - Database Languages - GQL.

  • Benefit from job function-based experiences:

    • Data engineers can model and create graphs.
    • Analysts can run low-code or no-code queries and curate view sets.
    • Business users can explore visually or use natural language to interact with the data.

  • Operate within Fabric: Automatically shut down when not in use and monitor usage in the capacity metrics app - all governed by Fabric OneLake security, compliance, and permission model.

Integration with Microsoft Fabric

Graph is deeply integrated with the Microsoft Fabric platform, including OneLake for unified data storage and Fabric UI for visualization. It integrates seamlessly with Microsoft Fabric's governance, security, and operational features.

You can incorporate graph analytics into your existing workflows, eliminating the need for data duplication and specialized skills. So, you can make insights accessible to a broader audience compared to traditional standalone graph databases.

How Graph differs from standalone graph databases

Area Graph Standalone graph database
Data gravity Graph operates directly on OneLake, so you don't have to perform ETL or duplicate data. Standalone graph databases require you to move or duplicate your data into a separate graph database instance, which can add complexity and overhead.
Scalability The service is designed for large-scale graphs and uses scale-out sharding across multiple machines to handle big data workloads efficiently. Most standalone graph databases rely on scale-up architectures or clusters that might be limited by the vendor or edition, which can restrict scalability.
Language Graph is compatible with the new GQL standard (preview) and includes built-in graph analytics algorithms. Standalone graph databases often use vendor-specific query languages and separate analytics frameworks. Support for algorithms can vary widely.
User experience Users benefit from a unified Microsoft Fabric interface for modeling, querying, business intelligence (BI), artificial intelligence (AI) integration, and low/no-code exploration. Specialized graph engineering skills aren't required. Standalone graph databases are primarily developer-focused, with consoles and SDKs that often require specialized skills. Visualization and low-code tools can be separate and might require extra setup.
Operations & cost Graph uses your existing Fabric capacity and automatically reduces resources when not in use, helping you save costs. Standalone graph databases require separate clusters or licenses, custom scaling and monitoring, and often incur idle capacity charges. They increase operational complexity and cost.
Governance & security Microsoft Fabric provides native OneLake governance, lineage tracking, and workspace role-based access control (RBAC). It integrates with Fabric compliance standards for security and auditing. Standalone graph databases have separate security and governance models that you must configure and audit independently. They can increase risk and administrative burden.

Note

Join the new Fabric user panel to share feedback and help shape Fabric and Power BI. Participate in surveys and one-on-one sessions with the product team. To learn more and sign up, see Fabric user panel.

Pricing and capacity units

Graph uses the same capacity units (CUs) as other workloads in Microsoft Fabric. You don't need to purchase a separate graph-specific license or SKU. All graph operations, including data ingestion, querying, and running algorithms, consume your organization's reserved or pay-as-you-go Fabric capacity.

Graph operations are billed based on CPU uptime. Each second of uptime incurs 10 CU-seconds in cost. Each session of CPU uptime is rounded up to minutes.

For graph storage, the system provisions a minimum of 100 GB. Graph storage is billed at the same rate as OneLake Cache.

For more information on pricing and capacity units, see Microsoft Fabric pricing.

You can monitor your graph workload's resource consumption and performance in the Fabric Capacity Metrics app. You see the following line items in the Fabric Metrics App and on monthly bills:

Fabric operation name in Metrics App Azure billing meter
Graph general operations Graph capacity usage CU
Graph cache storage OneLake Cache

Region availability

Graph is currently available in the following regions:

  • Australia East
  • Australia Southeast
  • Brazil South
  • Canada Central
  • Central India
  • Central US
  • East Asia
  • East US
  • East US 2
  • France Central
  • Germany West Central
  • Israel Central
  • Italy North
  • Japan East
  • Japan West
  • Korea Central
  • Mexico Central
  • North Central US
  • North Europe
  • Norway East
  • Poland Central
  • South Africa North
  • South Central US
  • Southeast Asia
  • South India
  • Spain Central
  • Sweden Central
  • Switzerland North
  • Switzerland West
  • UAE North
  • UK South
  • UK West
  • West Europe
  • West US
  • West US 2
  • West US 3

Related content

  • Last updated on