Summary
v1.124.0 adds Terraform resources and data sources for three AI Gateway components: model service, model provider service, and MCP service. This enables infrastructure-as-code management of AI Gateway deployments through Terraform.
Summary generated by brickster.ai. For the full changelog and any code/binary attachments, follow the GitHub link above.
More from databricks/terraform-provider-databricks
This release adds resources and data sources for private network gateways alongside Mason sessions and managed memory components. The resource exporter expands dependency support to include Lakeflow pipeline settings, model serving endpoints, Databricks Apps, cross-cloud cluster policies, and several new job task types.
The databricks_connection resource now supports schema-level connections through a new parent argument. The exporter fixes a nil pointer panic encountered when exporting workspace-level Unity Catalog securables with an account-level provider.
The databricks_secret resource now includes new string_value_wo and string_value_wo_version attributes. These additions allow users to manage write-only secret values within Databricks secret scopes.
Updated the Databricks API definitions to the latest version (#6020).
Fixed databricks_grant and databricks_grants for AI Gateway securables (model_service, mcp_service, model_provider_service), which previously failed when using resource names that included API prefixes like 'model-provider-services/'. The provider now automatically strips these prefixes before API calls and treats both prefixed and bare resource names as equivalent, eliminating apply failures and perpetual diffs.