Dagster sensor default_automation_condition_sensor stopped by default

[rafaljaronski]

Hi,
I'm trying to deploy Dagster on GKE via Helm charts with DBT models. Currently, I'm facing a problem with automated materialization.

To enable auto-materialization, I have tried a few approaches, with the most promising being:

• from DBT level:
  in dbt_project.yml i added:

    +meta:
         dagster:
           auto_materialize_policy:
             type: eager
  

  This creates a default_automation_condition_sensor, but it is disabled by default.
  (Another issue: for me, type: lazy doesn't do anything even after manually enabling the sensor.)
• from Dagster level:

  class CustomDagsterDbtTranslator(DagsterDbtTranslator):
    def get_auto_materialize_policy(self, dbt_resource_props: Mapping[str, Any]) -> Optional[AutoMaterializePolicy]:
        return AutoMaterializePolicy.eager().with_rules(AutoMaterializeRule.skip_on_not_all_parents_updated())
  
  @dbt_assets(manifest=dbt_manifest_path, dagster_dbt_translator=CustomDagsterDbtTranslator())
  def dbtbq_dbt_assets(context: AssetExecutionContext, dbt: DbtCliResource):
      yield from dbt.cli(["build","--profiles-dir","profile"], context=context).stream()
  

  This also creates a default_automation_condition_sensor, but it is disabled by default.
  However, it allows me to add the rule skip_on_not_all_parents_updated(), which I really want.

My question:
What is the simplest way to enable that sensor or change the default configuration?

I tried with:

# additionalInstanceConfig:
#   auto_materialize:
#     run_tags:
#       auto_materialized: "true"
#       trigger: "sensor"

but it doesnt work..