[ui] When viewing logs for a run, the date for a single log row is now shown in the tooltip on the timestamp. This helps when viewing a run that takes place over more than one date.
Added suggestions to the error message when selecting asset keys that do not exist as an upstream asset or in an AssetSelection.
Improved error messages when trying to materialize a subset of a multi-asset which cannot be subset.
[dagster-snowflake] dagster-snowflake now requires snowflake-connector-python>=3.4.0
[embedded-elt] @sling_assets accepts an optional name parameter for the underlying op
[dagster-openai] dagster-openai library is now available.
[dagster-dbt] Added a new setting on DagsterDbtTranslatorSettings called enable_duplicate_source_asset_keys that allows users to set duplicate asset keys for their dbt sources. Thanks @hello-world-bfree!
Log messages in the Dagster daemon for unloadable sensors and schedules have been removed.
[ui] Search now uses a cache that persists across pageloads which should greatly improve search performance for very large orgs.
[ui] groups/code locations in the asset graph’s sidebar are now sorted alphabetically.
Fixed issue where the input/output schemas of configurable IOManagers could be ignored when providing explicit input / output run config.
Fixed an issue where enum values could not properly have a default value set in a ConfigurableResource.
Fixed an issue where graph-backed assets would sometimes lose user-provided descriptions due to a bug in internal copying.
[auto-materialize] Fixed an issue introduced in 1.6.7 where updates to ExternalAssets would be ignored when using AutoMaterializePolicies which depended on parent updates.
[asset checks] Fixed a bug with asset checks in step launchers.
[embedded-elt] Fix a bug when creating a SlingConnectionResource where a blank keyword argument would be emitted as an environment variable
[dagster-dbt] Fixed a bug where emitting events from dbt source freshness would cause an error.
[ui] Fixed a bug where using the “Terminate all runs” button with filters selected would not apply the filters to the action.
[ui] Fixed an issue where typing a search query into the search box before the search data was fetched would yield “No results” even after the data was fetched.
[asset checks] UI performance of asset checks related pages has been improved.
[dagster-dbt] The class DbtArtifacts has been added for managing the behavior of rebuilding the manifest during development but expecting a pre-built one in production.
Microsoft Teams is now supported for alerts. Documentation
A send sample alert button now exists on both the alert policies page and in the alert policies editor to make it easier to debug and configure alerts without having to wait for an event to kick them off.
Added a new run_retries.retry_on_op_or_asset_failures setting that can be set to false to make run retries only occur when there is an unexpected failure that crashes the run, allowing run-level retries to co-exist more naturally with op or asset retries. See the docs for more information.
dagster dev now sets the environment variable DAGSTER_IS_DEV_CLI allowing subprocesses to know that they were launched in a development context.
[ui] The Asset Checks page has been updated to show more information on the page itself rather than in a dialog.
[ui] Fixed an issue where the UI disallowed creating a dynamic partition if its name contained the “|” pipe character.
AssetSpec previously dropped the metadata and code_version fields, resulting in them not being attached to the corresponding asset. This has been fixed.
The new @multi_observable_source_asset decorator enables defining a set of assets that can be observed together with the same function.
[dagster-embedded-elt] New Asset Decorator @sling_assets and Resource SlingConnectionResource have been added for the [dagster-embedded-elt.sling](http://dagster-embedded-elt.sling) package. Deprecated build_sling_asset, SlingSourceConnection and SlingTargetConnection.
Added support for op-concurrency aware run dequeuing for the QueuedRunCoordinator.
dagster-polars has been added as an integration. Thanks @danielgafni!
[dagster-dbt] @dbt_assets now supports loading projects with semantic models.
[dagster-dbt] @dbt_assets now supports loading projects with model versions.
[dagster-dbt] get_asset_key_for_model now supports retrieving asset keys for seeds and snapshots. Thanks @aksestok!
[dagster-duckdb] The Dagster DuckDB integration supports DuckDB version 0.10.0.
[UPath I/O manager] If a non-partitioned asset is updated to have partitions, the file containing the non-partitioned asset data will be deleted when the partitioned asset is materialized, rather than raising an error.
Fixed an issue where creating a backfill of assets with dynamic partitions and a backfill policy would sometimes fail with an exception.
Fixed an issue with the type annotations on the @asset decorator causing a false positive in Pyright strict mode. Thanks @tylershunt!
[ui] On the asset graph, nodes are slightly wider allowing more text to be displayed, and group names are no longer truncated.
[ui] Fixed an issue where the groups in the asset graph would not update after an asset was switched between groups.
[dagster-k8s] Fixed an issue where setting the security_context field on the k8s_job_executor didn't correctly set the security context on the launched step pods. Thanks @krgn!
Observable source assets can now yield ObserveResults with no data_version.
You can now include FreshnessPolicys on observable source assets. These assets will be considered “Overdue” when the latest value for the “dagster/data_time” metadata value is older than what’s allowed by the freshness policy.
[ui] In Dagster Cloud, a new feature flag allows you to enable an overhauled asset overview page with a high-level stakeholder view of the asset’s health, properties, and column schema.
[kubernetes] Fixed an issue where the Kubernetes agent would sometimes leave dangling kubernetes services if the agent was interrupted during the middle of being terminated.
Within a backfill or within auto-materialize, when submitting runs for partitions of the same assets, runs are now submitted in lexicographical order of partition key, instead of in an unpredictable order.
[dagster-k8s] Include k8s pod debug info in run worker failure messages.
[dagster-dbt] Events emitted by DbtCliResource now include metadata from the dbt adapter response. This includes fields like rows_affected, query_id from the Snowflake adapter, or bytes_processed from the BigQuery adapter.
A previous change prevented asset backfills from grouping multiple assets into the same run when using BackfillPolicies under certain conditions. While the backfills would still execute in the proper order, this could lead to more individual runs than necessary. This has been fixed.
[dagster-k8s] Fixed an issue introduced in the 1.6.4 release where upgrading the Helm chart without upgrading the Dagster version used by user code caused failures in jobs using the k8s_job_executor.
[instigator-tick-logs] Fixed an issue where invoking context.log.exception in a sensor or schedule did not properly capture exception information.
[asset-checks] Fixed an issue where additional dependencies for dbt tests modeled as Dagster asset checks were not properly being deduplicated.
[dagster-dbt] Fixed an issue where dbt model, seed, or snapshot names with periods were not supported.
@observable_source_asset-decorated functions can now return an ObserveResult. This allows including metadata on the observation, in addition to a data version. This is currently only supported for non-partitioned assets.
[auto-materialize] A new AutoMaterializeRule.skip_on_not_all_parents_updated_since_cron class allows you to construct AutoMaterializePolicys which wait for all parents to be updated after the latest tick of a given cron schedule.
[Global op/asset concurrency] Ops and assets now take run priority into account when claiming global op/asset concurrency slots.
build_schedule_from_partitioned_job now supports creating a schedule from a static-partitioned job (Thanks @craustin!)
[dagster-pipes] PipesK8sClient will now autodetect the namespace when using in-cluster config. (Thanks @aignas!)
[dagster-pipes] PipesK8sClient can now inject the context in to multiple containers. (Thanks @aignas!)
[dagster-snowflake] The Snowflake Pandas I/O manager now uses the write_pandas method to load Pandas DataFrames in Snowflake. To support this change, the database connector was switched from SqlDbConnection to SnowflakeConnection .
[ui] On the overview sensors page you can now filter sensors by type.
[dagster-deltalake-polars] Added LazyFrame support (Thanks @ion-elgreco!)
[dagster-dbt] When using @dbt_assets and multiple dbt resources produce the same AssetKey, we now display an exception message that highlights the file paths of the misconfigured dbt resources in your dbt project.
[dagster-k8s] The debug info reported upon failure has been improved to include additional information from the Job. (Thanks @jblawatt!)
[dagster-k8s] Changed the Dagster Helm chart to apply automountServiceAccountToken: false to the default service account used by the Helm chart, in order to better comply with security policies. (Thanks @MattyKuzyk!)
A unnecessary thread lock has been removed from the sensor daemon. This should improve sensor throughput for users with many sensors who have enabled threading.
Retry from failure behavior has been improved for cases where dynamic steps were interrupted.
Previously, when backfilling a set of assets which shared a BackfillPolicy and PartitionsDefinition, but had a non-default partition mapping between them, a run for the downstream asset could be launched at the same time as a separate run for the upstream asset, resulting in inconsistent partition ordering. Now, the downstream asset will only execute after the parents complete. (Thanks @ruizh22!)
Previously, asset backfills would raise an exception if the code server became unreachable mid-iteration. Now, the backfill will pause until the next evaluation.
Fixed a bug that was causing ranged backfills over dynamically partitioned assets to fail.
[dagster-pipes] PipesK8sClient has improved handling for init containers and additional containers. (Thanks @aignas!)
Fixed the last_sensor_start_time property of the SensorEvaluationContext, which would get cleared on ticks after the first tick after the sensor starts.
[dagster-mysql] Fixed the optional dagster instance migrate --bigint-migration, which caused some operational errors on mysql storages.
[dagster-dbt] Fixed a bug introduced in 1.6.3 that caused errors when ingesting asset checks with multiple dependencies.
[asset checks] @asset_check now has a blocking parameter. When this is enabled, if the check fails with severity ERROR then any downstream assets in the same run won’t execute.
Agent tokens can now be locked down to particular deployments. Agents will not be able to run any jobs scheduled for deployments that they are not permitted to access. By default, agent tokens have access to all deployments in an organization. Use the Edit button next to an agent token on the Tokens tab in Org Settings to configure permissions for a particular token. You must be an Organization Admin to edit agent token permissions.
Added support for the 3.0 release of the pendulum library, for Python versions 3.9 and higher.
Performance improvements when starting run worker processes or step worker processes for runs in code locations with a large number of jobs.
AllPartitionMapping now supports mapping to downstream partitions, enabling asset backfills with these dependencies. Thanks @craustin!
[asset checks][experimental] @asset_check has new fields additional_deps and additional_ins to allow dependencies on assets other than the asset being checked.
[ui] Asset graph group nodes now show status counts.
[dagster-snowflake] The Snowflake I/O Manager now has more specific error handling when a table doesn’t exist.
[ui][experimental] A new experimental UI for the auto-materialize history of a specific asset has been added. This view can be enabled under your user settings by setting “Use new asset auto-materialize history page”.
[ui] Command clicking on an asset group will now select or deselect all assets in that group.
[dagster-k8s] Added the ability to customize resource limits for initContainers used by Dagster system components in the Dagster Helm chart. Thanks @MattyKuzyk!
[dagster-k8s] Added the ability to specify additional containers and initContainers in code locations in the Helm chart. Thanks @craustin!
[dagster-k8s] Explicitly listed the set of RBAC permissions used by the agent Helm chart role instead of using a wildcard. Thanks @easontm!
Previously, calling get_partition_keys_not_in_subset on a BaseTimeWindowPartitionsSubset that targeted a partitions definition with no partitions (e.g. a future start date) would raise an error. Now, it returns an empty list.
Fixed issue which could cause invalid runs to be launched if a code location was updated during the course of an AMP evaluation.
Previously, some asset backfills raised an error when targeting multi-assets with internal asset dependencies. This has been fixed.
Previously, using the LocalComputeLogManager on Windows could result in errors relating to invalid paths. This has been resolved. Thanks @hainenber!
An outdated path in the contribution guide has been updated. Thanks @hainenber!
[ui] Previously an error was sometimes raised when attempting to create a dynamic partition within a multi-partitioned asset via the UI. This has been fixed.
[ui] The “Upstream materializations are missing” warning when launching a run has been expanded to warn about failed upstream materializations as well.
[ui] The community welcome modal now renders properly in dark mode and some elements of Asset and Op graphs have higher contrast in both themes.
[ui] Fixed dark mode colors for datepicker, error message, and op definition elements.
[ui] Pressing the arrow keys to navigate op/asset graphs while the layout is loading no longer causes errors.
[ui] Exporting asset and op graphs to SVG no longer fails when chrome extensions inject additional stylesheets into Dagster’s UI.
[ui] Dagster now defaults to UTC when the user’s default timezone cannot be identified, rather than crashing with a date formatting error.
[ui] Fixed an issue in the asset graph sidebar that caused groups to only list their first asset.
[ui] Fixed an issue where sensors runs would undercount the number of dynamic partition requests added or deleted if there were multiple requests for additions/deletions.
[docs] Fixed a typo in the “Using Dagster with Delta Lake” guide. Thanks @avriiil!
[asset checks] Fixed an issue which could cause errors when using asset checks with step launchers.
[dagster-webserver] A bug preventing WebSocket connections from establishing on python 3.11+ has been fixed.
[dagster-databricks] DatabricksJobRunner now ensures the correctdatabricks-sdk is installed. Thanks @zyd14!
[dagster-dbt] On run termination, an interrupt signal is now correctly forwarded to any in-progress dbt subprocesses.
[dagster-dbt] Descriptions for dbt tests ingested as asset checks can now be populated using the config.meta.description. Thanks @CapitanHeMo!
[dagster-dbt] Previously, the error message displayed when no dbt profiles information was found would display an incorrect path. This has been fixed. Thanks @zoltanctoth!
[dagster-k8s] PipesK8sClient can now correctly handle load_incluster_config . Thanks @aignas!
Added a new category to Concepts:Automation. This page provides a high-level overview of the various ways Dagster allows you run data pipelines without manual intervention.
Moved several concept pages under Concepts > Automation: Schedules, Sensors, Asset Sensors, and Auto-materialize Policies.
Fixed an issue where configuring the agent_queue key in a dagster_cloud.yaml file incorrectly failed to validate when using the dagster-cloud ci init or dagster-cloud ci check commands during CI/CD.
The warning for unloadable sensors and schedules in the Dagster UI has now been removed.
When viewing an individual sensor or schedule, we now provide a button to reset the status of the sensor or schedule back to its default status as defined in code.
Added experimental functionality which hides user code errors from the Dagster UI. You may enable this functionality by setting the DAGSTER_REDACT_USER_CODE_ERRORS environment variable to 1.
[dagster-dbt] @dbt_assets now accepts a required_resource_keys argument.
Fixed a bug where a run that targets no steps is launched by an asset backfill when code updates are pushed after backfill launch time.
Previously a graphQL error would be thrown on the asset backfill page if certain unpartitioned assets were changed to a partitioned assets. This has been fixed.
[ui] Show run log timestamps in the user’s preferred hour cycle (12/24h) format.
[ui] The “Export to SVG” option now works as expected in the improved asset graph UI.
[ui] On the asset graph, hovering over a collapsed group or the title bar of an expanded group highlights all edges in/out of the group.
Fixed occasional CI/CD errors when building documentation on a feature branch
Asset lineage graph UI revamp, to make it easier to visualize and navigate large graphs
Lineage now flows left-to-right instead of top-to-bottom.
You can expand and collapse asset groups in the graph.
A new left-hand sidebar provides a list of assets, organized by asset group and code location.
You can right-click on assets or groups to filter or materialize them.
You can filter by compute kind.
Dark mode for the Dagster UI – By default, Dagster will match your system’s light or dark theme but you can adjust this in the user settings in the top right of the UI.
Report asset materializations from the UI – I.e. you record an asset materialization event without executing the code to materialize the asset. This is useful in cases where you overwrote data outside of Dagster, and you want Dagster to know about it and represent it in the UI. It’s also useful when you have a preexisting partitioned asset and start managing it with Dagster: you want Dagster to show the historical partitions as materialized instead of missing.
MaterializeResult, AssetSpec, and AssetDep now marked stable – These APIs, introduced in Dagster 1.5, were previously marked experimental. They offer a more straightforward way of defining assets when you don’t want to use I/O managers.
Backfill previews – When launching a backfill that covers assets with different partitions, can you now click “Preview” to see that partitions for each asset that will be covered by the backfill.
Viewing logs for a sensor or schedule tick is no longer considered experimental – previously, accessing this functionality required turning on a feature flag in user settings.
Runs triggered by a sensor or schedule link to the tick that triggered them.
Default op/asset concurrency limits are now configurable at the deployment level, using the concurrency > default_op_concurrency_limit configuration in your dagster.yaml (OSS) or Deployment Settings page (Dagster Cloud). In OSS, this feature first requires a storage migration (e.g. dagster instance migrate).
Zero-value op/asset concurrency limits are now supported. In OSS, this feature first requires a storage migration (e.g. dagster instance migrate).
When a Nothing-typed output is returned from an asset or op, the handle_output function of the I/O manager will no longer be called. Users of most Dagster-maintained I/O managers will see no behavioral changes, but users of the In-Memory I/O manager, or custom I/O managers that store Nothing-typed outputs should reference the migration guide for more information.
[ui] The updated asset graph is no longer behind an experimental flag. The new version features a searchable left sidebar, a horizontal DAG layout, context menus and collapsible groups!
Previously, if a code location was re-deployed with modified assets during an iteration of the asset daemon, empty auto-materialize runs could be produced. This has been fixed.
The CLI command dagster asset materialize will now return a non-zero exit code upon failure.
[ui] The Dagster UI now shows resource descriptions as markdown instead of plain text.
[ui] Viewing stdout/stderr logs for steps emitting hundreds of thousands of messages is much more performant and does not render the Run page unusable.
[ui] Fixed an issue where sensors with intervals that were less than 30 seconds were shown with an interval of “~30s” in the UI. The correct interval is now shown.
[dagster-graphql] Fixed an issue where the GraphQL Python client raised an unclear error if the request failed due to a permissions error.
A slight change has been made to run status sensors cursor values for Dagster instance using the default SQLite storage implementation. If you are using the default SQLite storage and you are upgrading directly from a version of dagster<1.5.1, you may see the first tick of your run status sensor skip runs that completed but were not yet registered by the sensor during your upgrade. This should not be common, but to avoid any chance of that, you may consider an interim upgrade to dagster>=1.5.1,<1.6.0 first.
Added the ability to annotate code locations with custom agent queues, allowing you to route requests for code locations in a single deployment to different agents. For example, you can route requests for one code location to an agent running in an on-premise data center but requests for all other code locations to another agent running in the cloud. For more information, see the docs.