> in what situation is it really worth all the added complexity of risk of pushing out functions to the edge If you are talking about developer point of view,

> in what situation is it really worth all the added complexity of risk of pushing out functions to the edge If you are talking about developer point of view, then there is no additional complexity. All the complexity is covered by the underlying platform > what is the advantage of saving a few ms on a transaction? one example - if a transaction consists of a few separate sequential transactions, then ms add up and might affect user experience. Also an app might need to issue lots of requests on a page load and taken that there is a limit on parallel requests (6 requests per domain), the advantage might be sensible. Having said that, I tend to agree that many use cases are not sensible to a few ms advantage