so far no errors/missing data with the “working” 
approach - eg. dispatching through pubsub and centralizing the upsert in a single genserver; but am really wondering why such behaviour happens (the db timeout when each genserver was doing its upsert)
