Hey there we are seeing the following error on the tickers: sqlalchemy.exc.ArgumentError Running the jobs manually seems to work. Any suggestions on how to fix. We have tried restarting the daemon and dagit so that it picks up the latest image with no success.

Hi Pablo - if you add a "sqlalchemy<2" pin or upgrade to the latest version of dagster i suspect this will go away - a breaking change came in the latest version of the sqlalchemy package

Hello Daniel, Thanks for the response. We indeed update all our images / helm chart version to 1.1.14 (which is the latest), restart all daemon / pods but still no luck on the scheduled jobs 😞 sqlalchemy.exc.ArgumentError: Column expression, FROM clause, or other columns clause element expected, got [Column(‘id’, Integer(), table=<runs>, primary_key=True, nullable=False), Column(‘run_body’, Text(), table=<runs>), Column(‘status’, String(length=63), table=<runs>), Column(‘create_timestamp’, DateTime(), table=<runs>,

I see - can you share where exactly you're seeing that log message?

sure

Got it - so that error is happening in your user code deployment image. Are you certain that that image is on dagster 1.1.14 or has sqlalchemy pinned to < 2.0.0.?

yep so we explicitly use 1.1.14 on our 1. Helm chart version (checked that each pod has this ver) 2. Pull image dagster-k8s version with 1.1.14 (checked it downloads this ver) 3. Our custom image downloads (pip install) 1.1.14 for all dagster

Any chance you could share the Dockerfile for your user code deployment?

yeap here is the docker code image. So before updating below docker image, we couldn’t run the manual launch on Dagster website. After deploying this docker update, we were able to launch the job, but it fails at scheduled run

Is there any chance that there's a different version of dagster referenced in that last line that installed

dagster_module/.

? Does that have a setup.py?

thanks for the suggestion. let me quickly check on this

I'm not sure what to make of that - that's an error that only pops up for people on SqlAlchemy 2, every other report I've seen was fixed when they downgraded back to 1.4.46

let me double check on our entire flow.