---
title: "Harvest Metadata"
canonical: "https://docs.vaultspeed.com/space/VPD/3011510326/Harvest%20Metadata"
format: markdown
---
One of the key features of VaultSpeed is its ability to automatically harvest the metadata of your sources. This allows you to gain an immediate and thorough view of your data model, enabling you to efficiently prepare it to align with your data vault model.

The automated metadata harvesting process is facilitated through the VaultSpeed Agent, which supports various source types, including:

> Macro (children)

VaultSpeed also supports **key renaming automation** to help standardize your metadata for faster onboarding.

> ℹ️ ## Table Harvesting Limitation
> ℹ️ 
> ℹ️ A limitation on table harvesting within VaultSpeed has been introduced to optimize overall system performance. 
> ℹ️ 
> ℹ️ The **maximum limit for table harvesting is set at 30,000 tables**. This adjustment ensures the efficient functioning of the system.

> 📝 ## Source-Specific Limitations
> 📝 
> 📝 ### Snowflake Driver Limitation
> 📝 
> 📝 **Note**: There is a specific limitation related to the **Snowflake JDBC driver** used during the table harvesting process. Snowflake limits the number of tables returned when performing a `SHOW TABLES` query. This constraint is imposed by Snowflake itself and does **not originate from VaultSpeed software**.
> 📝 
> 📝 For more information, refer to the [official Snowflake documentation](https://docs.snowflake.com/en/sql-reference/sql/show-tables#:~:text=To%20view%20results%20for%20which%20more%20than%2010K%20records%20exist%2C%20either%20include%20LIMIT%20rows%20or%20query%20the%20corresponding%20view%20in%20the%20Snowflake%20Information%20Schema).
> 📝 
> 📝 ### Microsoft Fabric Limitation
> 📝 
> 📝 VaultSpeed supports metadata harvesting from Microsoft Fabric; however, **foreign key constraints cannot be harvested from Fabric sources using the Microsoft JDBC driver**, which does not expose foreign key metadata programmatically.
> 📝 
> 📝 This is consistent with behavior in **Azure Synapse**, where foreign keys are also not harvested due to similar limitations in Microsoft’s JDBC-based connectivity.
> 📝 
> 📝 Because of this:
> 📝 
> 📝 - **Foreign keys will not be automatically retrieved** during harvesting.
> 📝 - Users must **manually define foreign key relationships** in VaultSpeed if needed for modeling purposes.
> 📝 
> 📝 **Note:** This is a **driver-level limitation** imposed by Microsoft, **not a VaultSpeed defect**.