I’m trying to understand Fly pricing and finding it really stressful because I just can’t get my head around how the global deployment story actually works. According to your speedrun docs, default everything → app running globally. So I get the impression from that that Fly by default runs instances in every region, or spins up an instance in every region in response to traffic (like AWS Lambda, sort of?). I assume it’s more like the latter because presumably most apps aren’t getting traffic in every single region.
But then how does that correspond to your compute pricing? If you’re running on a
dedicated-cpu-1x for instance, the pricing page says that’s $31/mo for just one instance. Does that become $62/mo if you have users connecting from two regions, since Fly will be running that app in both regions? Is there some sort of prorating where you’re billed extra for the time the app is taking requests from both regions, and then it goes back if the app is only taking requests from one region? When you deploy an app you’re asked to choose a region, is it always running in that region and then also running in an arbitrary number of other regions?