How to keep dbt asset materialization for successfully materialized assets for failed runs?

[nixent]

Assume dbt asset consisting of two models A->B. If A is successful and B fails, whole run is registered as failed and A materialization is marked as failed as well. If raise_on_error=False then A will have materialization record, B - will not and whole run will be successful.
As a workaround, it is possible to handle it by sending an alert on asset failure or have outdated asset notification based on freshness policy

@dbt_assets(manifest=Path("target", "manifest.json"))
def my_dbt_assets(context: AssetExecutionContext, dbt: DbtCliResource):
   dbt_run_invocation = dbt.cli(["run"], context=context, raise_on_error=False)

   if dbt_run_invocation.is_successful():
       yield from dbt_run_invocation.stream()
   else:
      send_asset_failure_alert(...)

I'm looking for a way to retain dbt asset materialization for successful assets and record run failure. In this case run status would be partially successful. This is also relevant for dbt tests with severity=warn

[rexledesma]

You could raise a Failure in the body of your computation.

@dbt_assets(manifest=Path("target", "manifest.json"))
def my_dbt_assets(context: AssetExecutionContext, dbt: DbtCliResource):
   dbt_run_invocation = dbt.cli(["run"], context=context, raise_on_error=False)

   yield from dbt_run_invocation.stream()
   
   if not dbt_run_invocation.is_successful():
       raise Failure(description="partial success")