is there currently a way (or a way in the works) for marking a run as failed if the dbt cli returns an error? I had a model repeatedly failing, but didn't notice it because the runs would always succeed

These will now fail once https://github.com/dagster-io/dagster/pull/15055 is released. We didn’t have a release last week, so it will be released this week. There’s a thread of feedback from the last time this was encountered this, as well as in the discussion, and looks like this behavior is unintuitive. So we’re defaulting for this to fail the run, but it can be optionally suppressed.