I've looked at that and probably will do that for dev/testing. Is it worth it to stay there and start paying after you've proven yourself or is that the mark to move on?
Good question. GAE isn't without problems of its own, but at first glance looks good enough to scale to millions of users. I've never had a project that has done that, but it seems the time to move on is when you run into issues with it that you cannot overcome.