question

[Update for clarification of the AWS billing model - and putting that up top, as it's importnat to understand]

For custom game server usage, we bill the actual cost from EC2 plus 10%, so their "simple" monthly calculator can get you more accurate cost modeling, if necessary: https://aws.amazon.com/blogs/aws/estimate-your-c/.

The two things to know, aside from the type (On-Demand), model (t2.medium by default, but once you've set up billing, let us know if you want something else), and OS (Windows), are:

1. EC2 billing is largely CPU hours (that's normally the majority of the cost), but it also includes data transfer costs. The calculator gives you the ability to model that out, but unless you're a heavy data user, it's not going to add much.2. The specific cost I see on the t2.medium line on the On-Demand Windows server (as of today, 2017.05.23) is $0.065 per hour. The thing to be aware of is that they specifically charge per hour. No smaller increments. So if you spin up a server and let it run for one minute or fifty nine (I'd say sixty, but trying to run a server for exactly sixty minutes would be an exercise in gambling), then shut it down, they bill for an hour either way. The next server you start is going to start a new hour. Another example would be that if you start two servers and run them both for a minute then shut them down, that would be billed to us as two hours - one per server. So for managing costs, it's important to schedule your server testing so that you maximize your time.Specific to the original question though, the running server hours increments while you have active custom game server hosts running, and those servers cost money to run, yes.

Basically, we provide a small number of hours in accounts to start, so that developers can do some testing. Since the way our server hosting works is that we automatically spin up additional servers if you don't have enough running to have the min free slots from your build configuration available, you could have multiple servers running, each consuming some of that time. We'll be providing a way to set your "max server hosts" to something less than 10 (the default) in the limits tab shortly.

For now, the first thing to do is to set your "Max instances per host" and "Min free slots" to be accurate. The max instances per host is how many instances of your exe can run on a host machine, while the min free slots is how many instance slots need to be available to run immediately (so that players trying to join games don't have to wait on a server host to start, as that takes time).

Once you've done that, the next thing is to only activate servers in regions when you want to test, or when you're going live. Setting a region active means you're requesting servers to run in each of those regions.

And finally, if you need more hours, you'll need to have billing set up for your title - either credit card info, or else via a contract with us that specified invoice billing. Once you have that set up, just let us know what to set the max hours to for your titles (as well as what server model you want to use), and we'll get that configured for you. The cost for the servers is then specifically the EC2 cost (https://aws.amazon.com/ec2/pricing/on-demand/ - the Windows tab) plus 10%.