Automation condition sensors

All automation conditions must be evaluated by a sensor. If you have any assets with an automation condition in your code location, a sensor with the name default_automation_condition_sensor will be created for you automatically. You'll need to toggle this sensor on in the UI for your code location under the Automation tab.

By default, this sensor will evaluate all automation conditions in your code location, and execute the asset or asset check if the condition is met.

The default_automation_condition_sensor is created when all of the following are true:

• The code location contains at least one asset with an AutomationCondition (or legacy AutoMaterializePolicy).
• You're on Dagster 1.5 or later.
• The code location loads and registers definitions successfully.

If you're migrating from AutoMaterializePolicy, you must set use_sensors: true in your auto_materialize configuration. Toggling sensors on in the UI alone is not enough; without use_sensors: true, the sensor will not drive automation. If a default sensor is missing, verify the conditions above before troubleshooting further.

Adding additional sensors

If you have a code location with a large number of assets using automation conditions, you may want to create additional sensors to evaluate only a subset of the conditions. This has a few benefits:

• Individual evaluations will execute more quickly
• Operational issues with a single sensor will not impact unrelated assets managed by a different sensor
• You can toggle automation on and off for targeted sets of assets

To add an additional sensor, create a AutomationConditionSensorDefinition and pass it to the sensors argument of your Definitions object:

src/<project_name>/defs/sensors.py

import dagster as dg


@dg.definitions
def sensors():
    return dg.Definitions(
        sensors=[
            dg.AutomationConditionSensorDefinition(
                "team_b_sensor", target=dg.AssetSelection.groups("team_a")
            ),
            dg.AutomationConditionSensorDefinition(
                "team_a_sensor", target=dg.AssetSelection.groups("team_b")
            ),
        ],
    )

When you create new sensors, the default_automation_condition_sensor will only target the automation conditions that are not already covered by the additional sensors.

Default to running

As with other sensor types, all AutomationConditionSensorDefinitions will default to the STOPPED state when they are first added to a code location. To modify this behavior, create an AutomationConditionSensorDefinition with the default_status argument set to RUNNING:

src/<project_name>/defs/sensors.py

import dagster as dg


@dg.definitions
def sensors():
    return dg.Definitions(
        sensors=[
            dg.AutomationConditionSensorDefinition(
                "run_tags_automation_condition_sensor",
                target=dg.AssetSelection.all(),
                default_status=dg.DefaultSensorStatus.RUNNING,
            )
        ],
    )

Adding run tags

To add tags to runs produced by a given AutomationConditionSensorDefinition, pass a run_tags argument to the AutomationConditionSensorDefinition:

src/<project_name>/defs/sensors.py

import dagster as dg


@dg.definitions
def sensors():
    return dg.Definitions(
        sensors=[
            dg.AutomationConditionSensorDefinition(
                "run_tags_automation_condition_sensor",
                target=dg.AssetSelection.all(),
                run_tags={"key": "value"},
            )
        ],
    )