Without making any promises here, I believe you can achieve somewhat similar functionality using Prefect 2.0 with
work queues.
We don't directly have support for Celery, but as mentioned in this linked issue, if you would want to build that, we would help to review and merge such a new execution backend. But given that Prefect 2.0 is the "LTS product", it would be interesting to see how you could implement it with
Orion and together with work queues.