Hi Edmond, I'm using it on Databricks, where it is an MLflow is an integrated managed service. It has been a smooth experience so far. However since most wouldn't have access to managed Databricks, I demonstrated the usage locally, in this article.
That being said I came across this resource which uses Amazon EC2 and Amazon RDS (both on the free tier) for deploying MLflow. Might be worth checking out. https://allcloud.io/blog/organise-your-ml-experiments-with-mlflow-on-aws/
Hope this helps!