databricks Databricks Release Notes ·

Databricks Unity Gateway Now Monitors and Caps External Model Provider Costs

aiazuredatabrickspreviewengineer
feature

Databricks has enhanced Unity Gateway to track and cap spend on external model providers like Amazon Bedrock and Azure AI Foundry, extending beyond Databricks-provided models. This new capability positions Unity Gateway as a single control point for managing usage and enforcing spend limits across teams, particularly beneficial for organizations with existing capacity commitments. Users can enable this feature via the "External Model Spend" setting within Budgets (Beta) on the account console's Previews page. It requires selecting "Include external model usage" when creating a budget for Unity Gateway resources.

Features (1)
  • Monitor and Cap External Model Provider Costs in Unity Gateway

    Unity Gateway now supports tracking and capping spend for external model providers such as Amazon Bedrock and Azure AI Foundry. This allows organizations with direct provider commitments to manage usage and enforce spend limits through a single interface. The functionality is accessible by enabling "External Model Spend" in Budgets (Beta) on the account console's Previews page and selecting "Include external model usage" when configuring a budget for Unity Gateway resources.

Read the original announcement →

https://docs.databricks.com/aws/en/release-notes/unity-gateway#monitor-and-cap-external-provider-costs

Related releases