dagster-dbt's fetch_row_counts() doesn't work and prints my MotherDuck connection string in the error logs

[rob-teeuwen]

I am seeing my MotherDuck connection string printed in a Dagster error log, which seems concerning.

Context: I wanted to turn on row counts meta data tracking for my dbt models as per docs here. So I’m passing that function fetch_row_counts(). It doesn’t work, because for some reason, it’s trying to open a file and the filename is too long. It prints that filename in an error message, and interestingly it includes my entire connection string. This is what the error looks like in Dagster Cloud.

  IO Error: Cannot open file "/venvs/13541fde3740/lib/python3.11/
site-packages/working_directory/root/dbt-project/md:production?
motherduck_token=[MY MOTHERDUCK TOKEN WAS HERE]": File 
name too long
ERROR:dagster.builtin:materialize_dbt_order_models - 2a9e2e92-
148e-4f2e-a154-fe62556c2b3f - gardyn_dbt_assets - An error 
occurred while fetching row count for model.gardyn_dbt.base_gardyn
__membership_ledger. Row count metadata will not be included 
in the event.

When I’m running this locally I get the same error but in addition I can verify that there is no such file in the directory where it points (it just points to the root of my dbt repo on my local disk, and it obviously doesn’t have a file with just my connection string and no extension).

Everything else about my dbt-dagster integration works great, and has done so for the past year or so. I haven't changed anything in the configuration (but I'm on the latest version so some of my config may have gotten outdated since I haven't changed anything).

I'm defining my dbt assets like this: (I'm importing the correct manifest path based on some env variables, this is how it used to be described in the docs. Now it uses DbtProject which seems to do the same thing as I am doing manually).

@dbt_assets(manifest=dbt_manifest_path)
def gardyn_dbt_assets(context: AssetExecutionContext, dbt: DbtCliResource):
    yield from dbt.cli(["build", f"--target={dbt_target}"], context=context).stream().fetch_row_counts()