I need some clarification on how the 3 shared-cpu-1x 256mb VMs limitation works in practice.
Does it mean one can only have 3 such VMs existing at any point in time?
Does it mean one can have even 10 VMs created, but only 3 can be running at any given time?
The scenario I have in mind is one where I have several projects deployed,
yet no more than 3 of them are ever running at the same time
(due to projects scaling down to 0 machines when inactive).
My understanding is that you essentially get a $5 allowance each month. As long as your total usage is under that, you don’t have to pay anything.
From their docs:
Fly Machines are fast to start and stop, and you don’t pay for their CPU and RAM when they’re in the stopped state.
So … if you only ever had 7 stopped machines and 3 running machines at any point in time, you would only incur costs from the three running ones. Assuming that total usage for the month was under $5, that’s free. It would be as if the 7 stopped machines didn’t exist at all.
But … if those X apps have other provisioned resources (ie beyond CPU and RAM), for example an IPv4 address, or XGB of SSD storage, I assume those resources would remain provisioned despite the machine being stopped. And so would incur costs. As they are not available for other customers to use (unlike CPU and RAM). If your apps need additional resources like those, you may exceed the free allowance via that route.