More than one instance being created per scale evaluation

Using the following scaling policy:

when the ScaleUp policy is triggered, it creates more than once instance (presumably because the evaluation period elapses multiple times before the new server has the chance to register that more slots are now available?)

Is this expected behavior? Is the evaluation period too short? A longer evaluation period would add to the already long server instance spin-up time (since we’re currently on a Windows isntance).

One other way I’m open to try is using an absolute scaling method (Scale to X if Current player sessions >= Y). Would using multiple of those conflict? Say, for instance, would:

[Scale to 2 if CurrentPlayerSessions >= 12]

[Scale to 3 if CurrentPlayerSessions >= 28]

[Scale to 4 if CurrentPlayerSessions >= 44]

conflict in a way that it continuously scales down to 2 if above 44 players?

It’s easy enough for me to generate those ranges with the script I have calling the AWS CLI, but again, there’d be up to 50 of them (our currentl instance limit in some regions)

Hey @funbits_cjacobson -

You should only be getting one “round” of instances per scaling evaluation. When you look at your fleet events log, do you see multiple scaling events in close succession or spaced apart 5+ minutes? Another thing that could be happening - if you have multiple scaling policies that conflict, GameLift will execute the policy that makes the biggest change. So if you had two scale-up policies, both valid, with one requesting 1 instance and another 2 instances, 2 would be assigned. Looking at your screenshot though, looks like you just have one scale-up configured.

We’d like to dig in a bit further here. Could you provide us with the region and either a fleet or game session identifier so we can debug on our end?