Looks like this is the case, which might have caused app to regularly crash on AWS. Will try to enforce a single client that handle all DB operations