Edit

Share via

Limitations in Cosmos DB in Microsoft Fabric (preview)

  • Applies to: ✅ Cosmos DB in Fabric

Important

This feature is in preview.

This article lists current limitations for Cosmos DB in Fabric.

The content in this article changes regularly. For the latest limitations, revisit this article periodically.

Quotas and limits

  • Containers support a maximum autoscale throughput of 10,000 request units per second (RU/s).

  • Containers created in the Fabric portal are automatically allocated 5,000 RU/s. Containers created using a software development kit (SDK) can be set up to the maximum allowed autoscale throughput.

  • Container throughput can't be changed once the container is created.

Regional availability

  • The following regions aren't supported as Cosmos DB in Fabric regions. These regions can be your home region, but can't contain Fabric capacity used for Cosmos DB in Fabric:

    • West US 3

    • Central US

    • South Central US

    • Switzerland North

Data

  • JSON strings greater than 8 kB are truncated when queried from the mirrored SQL analytics endpoint. The query editor includes the following error message:

    JSON text is not properly formatted. Unexpected character '"' is found at position
    • This limitation is related to a similar limitation of the data warehouse feature. The current workaround is to create a shortcut of the mirrored database in Fabric Lakehouse and utilize a Spark notebook to query your data.

  • quantizedFlat and diskANN indexes require at least 1,000 vectors to be indexed to ensure that the quantization is accurate. If fewer than 1,000 vectors are indexed, then a full-scan is used instead and RU charges might be higher.

  • Vectors indexed with the flat index type can be at most 505 dimensions. Vectors indexed with the quantizedFlat or DiskANN index type can be at most 4,096 dimensions.

  • The rate of vector insertions should be limited. Large ingestion, in excess of 5,000,000 vectors, could require extra index build time.

  • The vector search feature isn't currently supported on the existing containers. To use it, a new container must be created, and the container-level vector embedding policy must be specified.

  • Shared throughput databases are unsupported.

  • At this time, vector indexing and vector search aren't supported on accounts with Analytical Store, Azure Synapse Link, or shared throughput.

  • Once vector indexing and vector search are enabled on a container, they can't be disabled.

Full text indexing

  • Multi-language support is only supported with the following languages:

    • en-US (English)

    • de-DE (German)

    • es-ES (Spanish)

    • fr-FR (French)

Authorization

Programmability

  • Stored procedures, triggers, or user-defined functions aren't supported.

Related content