SDKs and Terraform added IAM v2 management APIs and schema updates for service principals.

August brought widespread updates for managing service principal identities across Databricks development tooling. Multiple SDK releases and the Terraform provider added support for IAM v2 APIs at both the account and workspace levels [2], [3], [4], [5]. These updates introduced full lifecycle operations to create, read, list, and delete service principals, alongside corresponding methods for groups, users, and workspace assignments [2], [3], [5].

The transition to newer identity APIs also enforced stricter schema requirements. Successive SDK releases introduced breaking changes that designated several service principal and user attributes as required fields [6], [7]. Developers and platform engineers maintaining automation scripts had to update their API payloads to include these mandatory attributes when upgrading SDK versions [6], [7].

At the provider level, version 1.125.0 resolved a recurring friction point for automated deployments [8]. Workspace-level hosts began resolving workspace identifiers directly from host metadata instead of issuing SCIM calls [8]. This design prevented authentication failures for non-interactive service principals configured without access to the SCIM /Me endpoint, while also catching workspace ID mismatches during plan time rather than execution time [8]. Meanwhile, community discussions tracked the arrival of serverless access control, focusing on how permissions map to automation workloads [1].