Table of Contents

Data Sources

Learn how to configure and manage data sources in the Management Portal.

Overview

Data sources define the storage locations from which agents and data pipelines can retrieve information. They provide the connection configuration needed to access your organizational data.

Accessing Data Sources

  1. In the Management Portal sidebar, click Data Sources under the Data section
  2. The data sources list loads, showing all configured sources

Data Source List

The table displays:

Column Description
Name Unique identifier for the data source
Source Type Type of storage (Azure Data Lake, SharePoint, etc.)
Edit Settings icon to modify configuration
Delete Trash icon to remove the data source

Searching and Managing

  • Use the search box to filter by name
  • Click the refresh button to reload the list
  • Click column headers to sort

Supported Data Source Types

Type Description
Azure Data Lake Azure Data Lake Storage Gen2 containers
OneLake Microsoft Fabric OneLake storage
SharePoint Online SharePoint document libraries
Azure SQL Database Azure SQL databases
Web Public web content via URLs

Creating a Data Source

  1. Click Create Data Source at the top right of the page
  2. Configure the data source settings

Basic Information

Field Description Requirements
Data Source Name Unique identifier Letters, numbers, dashes, underscores only
Description Purpose and contents Optional but recommended

Select Data Source Type

Choose the appropriate type from the dropdown:

  • Azure Data Lake
  • OneLake
  • SharePoint Online
  • Azure SQL Database
  • Web

Configure Connection Details

Connection settings vary by data source type:

Azure Data Lake

Field Description
Authentication Type How to authenticate (Connection String, Account Key, Azure Identity)
Connection String Full connection string (if using Connection String auth)
API Key Storage account key (if using Account Key auth)
Endpoint Storage account endpoint URL
Account Name Storage account name (if using Azure Identity)
Folders Container/folder paths to access (press Enter after each)

Authentication Options:

Type Description Best For
Connection String Full connection string with SAS or key Development, testing
Account Key Storage account access key Simple deployments
Azure Identity Managed Identity authentication Production, secure access

OneLake

Similar to Azure Data Lake, with authentication options for Fabric workspaces.

SharePoint Online

Field Description
Site URL SharePoint site URL
Document Library Library to access
Authentication Microsoft Graph authentication settings

Azure SQL Database

Field Description
Server SQL Server hostname
Database Database name
Authentication SQL or Azure AD authentication
Tables/Views Specific tables or views to access

Web

Field Description
URL(s) Web pages to crawl
Crawl Depth How deep to follow links

Specifying Folders

For storage-based sources, specify folders using chips:

  1. Type the folder path (e.g., container/folder)
  2. Press Enter or , to add the chip
  3. Repeat for additional folders
  4. Click the X on a chip to remove it

Editing Data Sources

  1. Locate the data source in the list
  2. Click the Settings icon (⚙️)
  3. Modify settings as needed
  4. Click Save Changes

Note: The data source name cannot be changed after creation.

Deleting Data Sources

  1. Click the Trash icon (🗑️) for the data source
  2. Confirm deletion in the dialog

Warning: Deleting a data source affects any data pipelines or agents using it. Verify dependencies before deleting.

Access Control

Configure who can access and manage the data source:

  1. Open the data source for editing
  2. Click Access Control at the top right
  3. Add or remove role assignments
Permission Description
FoundationaLLM.DataSource/dataSources/read View the data source
FoundationaLLM.DataSource/dataSources/write Edit configuration
FoundationaLLM.DataSource/dataSources/delete Delete the data source

Best Practices

Security

  • Use Azure Identity (Managed Identity) for production environments
  • Avoid embedding secrets in connection strings when possible
  • Apply principle of least privilege to folder access

Organization

  • Use descriptive names that indicate the data content
  • Document the purpose in the description field
  • Group related data in logical folder structures

Performance

  • Limit folder scope to only necessary paths
  • Consider separate data sources for different data types

Troubleshooting

Connection Test Fails

  • Verify authentication credentials are correct
  • Check network connectivity to the storage endpoint
  • Ensure firewall rules allow access from FoundationaLLM services

Data Not Accessible

  • Verify folder paths are correct
  • Check that the service principal or managed identity has appropriate permissions
  • Review Azure Storage access policies