> At this point, some people are probably screaming at their screens going: "just use FastAPI!" -- and we did indeed consider it.

Working with both sync Django and async FastAPI daily, it’s so easy to screw up async FastAPI and bring things to a halt. If async is such the huge key feature they seem to think it is for their product, then I would agree moving away from Python early while it’s still relatively easy is the right call.

> and we had actually already written our background worker service in Node,

Ok well that’s a little bizarre… why use Django to begin with if you are not going to use the huge ecosystem that comes with it. New Django has first-class support for background workers, not that Celery is difficult to get setup. It’s sounds like the engineering team just started building things in what they knew without any real technical planning and the async hiccup is more or less an excuse to get things in order after the fact.

> it’s so easy to screw up async FastAPI and bring things to a halt

For example?