Having troubles with airflow.
Hey guys. Most of our stuff ran in cron before. And I decided to make things more reliable. So I setup self hosted airflow in docker etc. But it's been quite a pain. It keeps getting stuck every few days silently due to one or the other random reason every time.
I was using external python operator before inside the same docker as the scheduler. But then I it got stuck in hangups etc and I thought that's the issue so I did it in a more fancy way with 4-5 containers celery, redis, scheduler etc in separate containers. And even today it got stuck on one job randomly. I was on airflow 3.0.0 before though we upgraded it to 3.2.x or something today to see if that helps. But it's been a bit of a fight. That I am starting to get a bit tired.
I had hoped that it being the industry standard and all it would be super smooth a perfect but it's been a bit of a pain in the ass. I am not sure if it's airflow itself that's at fault or am I doing something wrong. I am not an airflow expert and working with ai on it. So I might be missing something. But it has not been a smooth experience and I am considering just using cron, or potentially dagster. But let me know what you guys think. Maybe a managed solution is better but I would like if it's something we can stay on free tier of. As it's a pretty shit dumb low reliability job that cron can almost take over with 0 reliability issues.
Let me know what you guys suggest and if I am doing something wrong. Thanks 🙏🏻