I’m still quite new to gamelift services and spending considerable time testing and learning at my own pace. I’d like to use my free tier hours for this learning stage but if I leave a fleet running for hours during which I’m not tinkering around, it would basically be eating away at those free hours… is there a way to avoid this without having to wait for a fleet to activate every time I want to do some actual online testing (not using gamelift local)? So far I tried to create a fleet of type spot and manually set the instances to 0, however I could still log into game sessions from the client. I think it had to do with having protection set to “full” but I wanted to get some solid direction on here before I get to trying this again.
If anyone can point me in the right direction I would appreciate it, perhaps there is an easier way through CLI commands as well?
Setting the instance count to 0 using the console or UpdateFleetCapacity API, is the correct approach for this.
You’re correct that this will not scale-down the fleet if it has an ACTIVE GameSession on it and protectionPolicy is set to
FullProtection. You can remove protection from a particular GameSession by calling the UpdateGameSession API or remove protection from all new GameSessions created on a fleet by using the UpdateFleetAttribute API
These APIs can be called via the AWS CLI - update-game-session — AWS CLI 1.22.37 Command Reference
One more thing to note, since you mentioned free-tier usage is that GameLift free-tier is only eligible for On-Demand instance usage. So make sure you’re not using SPOT if you wish to use those hours
Hi @Nathan ,
Thanks for the detailed explanation and all the helpful links! I was under the impression that I wouldn’t have any control over instance count for On-Demand instance usage.
Hey @wayne_grech ,
You are able to control instance counts for on-demand usage the same as spot usage (primarily by the defined fleet capacity).
The main difference between the two is that SPOT instances are cheaper, but are subject to the possibility of being interrupted (and shut-down) when AWS needs the capacity back. You can read more about the differences here - Choosing computing resources - Amazon GameLift
Thanks @Nathan that clears up the confusion!