Dagster 1.13.11 - GraphQL Counts and dbt Insights

Dagster 1.13.11, published on June 25, 2026, is a focused scheduler and data engineering release for larger asset graphs, dbt metadata, and GraphQL clients. The main operator change is bounded GraphQL selection data for runs, which helps API consumers show asset and check scope without reading more run data than they need. This release is not marked as a prerelease.

The full release notes and downloads are on the GitHub release page.

The dagster-graphql change is small, but it is the most useful one for platform teams that read Dagster state outside the main UI. The Run GraphQL type now has an optional limit argument and two new fields, assetSelectionCount and assetCheckSelectionCount.

That gives clients two separate pieces of information. A client can render a bounded preview of the assets and checks selected by a run, while also showing the true totals. This matters when a run covers a wide asset graph and the page, service, or report only needs enough detail for orientation.

The release notes do not describe a schema migration or a required client change. Still, any internal tool that queries Run for selection data should check whether it can replace larger reads with the new count fields. That is the practical value here: smaller GraphQL payloads and clearer totals in operator views.

Dagster 1.13.11 also changes how the asset catalog page behaves for large workspaces. The page now renders as a single virtualized list. The stated result is that workspaces with many asset groups or code locations should no longer freeze the UI when users expand those sections.

This is a UI change, but it has a data engineering impact. Large Dagster deployments often use asset groups and code locations as real operational boundaries. If expanding those boundaries blocks the browser, operators lose a basic navigation path through the graph. Virtualization is a sensible fix because it limits how much DOM work the browser does at once.

The run detail page header now displays the asset-check count for asset-job runs. It also reads that count from the execution plan instead of scanning the full run event log. That is a cleaner source for a summary value, and it avoids making the header depend on a larger event history read.

The dagster-dbt update adds an insights option to the include_metadata field on DbtProjectComponent. The point is specific: Dagster+ Insights tracking can now be enabled from YAML config.

For teams that keep dbt project wiring in declarative config, that is better than moving the same concern into code. It keeps the Insights choice near the rest of the project component settings, which is where operators will usually look during review.

There is also a safer failure mode for .with_insights(). When it is called against an unsupported adapter, it now logs a warning instead of raising. That does not make an unsupported adapter supported. It does make mixed adapter environments less brittle, because the failure is reported without turning the call into an exception.

Asset check history is now cleared when an asset or its partitions are wiped. The visible effect is that stale entries should no longer remain in the Execution History and Partitions views after the underlying asset data has been removed.

The run actions menu also gets a scope fix. The “View asset lineage” item now scopes to the assets in that run. For anyone debugging a run inside a dense graph, that should reduce stray context and make lineage views line up with the selected execution.

The documentation changes are also operator focused. Dagster added a page for Dagster+ code location snapshot size limits, including a local script for estimating snapshot sizes before deployment. The release also documents the “Run concurrency” Insights metric and fixes an invalid pool config in the “Preventing concurrent dbt snapshots” DbtProjectComponent example.

None of these notes are flagged as breaking changes. The main review items are GraphQL clients that can use the new count fields, dbt projects that want include_metadata set from YAML, and large workspaces that previously hit browser freezes in the asset catalog.

• Release page: GitHub release page
• Project repo: dagster-io/dagster
• Tag: 1.13.11