@Anna Geller I am using Prefect 1. And the old 1.x Prefect Cloud. The setup is Prefect flows are registered on AWS ECS with Fargate on 16vcpu machines.
When I run 4-6 flows concurrently, there is no issue. Once I pass 7-9 concurrent flows, some of them freeze. They still appear as alive but when i check the logs they are dead.
The machines are big enough to run them, the flows usually take abt 8-10 hrs at most and my aws account has enough vCpus available for fargate to run them