Oct 20, 2026: dbt Projects on Snowflake updates

The following improvements are now available for dbt Projects on Snowflake:

  • dbt State integration: Connect dbt project objects and Workspaces to dbt State, a dbt Labs service that decides which models need to be rebuilt and which models can be reused during an execution. Skipping models that don’t need to run aims to reduce execution runtime and warehouse compute. You connect your project to your dbt Platform account through an env.yml secret and an external access integration. For setup, supported dbt versions, and billing, see Integrating dbt State with dbt Projects on Snowflake.

  • Source freshness: Source freshness builds on the same dbt artifact support that powers Slim CI. Run source freshness from the command list in Workspaces, from the dbt project object details page, and when you schedule a dbt project object. Snowflake CLI also supports source freshness with snow dbt execute. After a freshness check, source nodes in the Workspaces DAG and the dbt project object details DAG show a Fresh, Expiring, or Stale badge. The badge is recalculated from the last load time and the warn and error thresholds in your source YAML, so it reflects the current state rather than only the last check. Hover over a badge to see when the data was last loaded and the freshness rule, or select the source node to see the freshness rule in the side panel. To retrieve freshness results with SQL, see Retrieve source freshness results.

  • Cross-project references in the DAG: When your project pulls in another dbt project as a private Git package with dbt deps, such as a subfolder of a monorepo or another team’s repository, and references its models with the two-argument ref('project', 'model') syntax, the DAG now makes those cross-project dependencies visible. In the Workspaces, Query History, and dbt project object details DAGs, nodes from the imported project are labeled with their source project, and the DAG shows only the imported models that your project references instead of every model in the package. A filter in the DAG toolbar isolates cross-project nodes and their direct neighbors, and a project topology view shows which dbt projects your project depends on. To set up private Git packages, see Admin setup. For more information, see Cross dbt project dependencies and Explore the DAG from a workspace.

  • Re-run failed models from Query History: In the dbt Output section of the query details for a failed dbt project object execution, select Re-run to rerun the models that errored and their downstream dependencies with --select result:error+. You can use the artifacts from that exact query or from the last failed run target, or run dbt retry. This option is available for dbt project objects that use the live version. For more information, see Recover from a failed execution.

  • dbt Projects home page and templates: The dbt Projects home page in Snowsight lists all of your dbt project objects, with filters for database, owner role, and name, and a Run History tab. From the home page, you can start a new project from a template, such as the Tasty Bytes sample project or a dbt with Cortex project that opens Cortex Code to help you build it. The Git-free Getting started with dbt Projects template is also available in the Snowsight templates gallery and from a tile on the Workspaces home page. It adds a ready-to-run Tasty Bytes dbt project to a workspace. For more information, see Quick start: Use the Snowsight app template.

  • Environment in run history: The dbt Projects run history in Snowsight shows the env.yml environment selected for each run. For more information, see Monitor dbt project objects in Snowsight.