Reputation: 20045
I have Fargate on a cluster running behind a NLB. I'd like to add autoscaling for it. But I can't find an example for that.
This is the example I found for an ApplicationLoadBalancedFargateService
:
fg = aws_ecs_patterns.ApplicationLoadBalancedFargateService(...)
scaling = fg.service.auto_scale_task_count(max_capacity=10)
scaling.scale_on_request_count(
id="RequestScaling",
requests_per_target=10000,
target_group=fg.target_group
)
But this won't work because the target_group
parameter expects an instance of ApplicationTargetGroup
but for NetworkLoadBalancedFargateService
the target_group
field is of type NetworkTargetGroup
.
Upvotes: 1
Views: 1041
Reputation: 2768
NLBs don't have a RequestCount metric type just because I guess each TCP connection could contain a number of "requests" in the ALB sense.
In the short term, I think the ActiveFlowCount matches closest with Request Count - so you can use the scaleOnMetric method to scale on ActiveFlowCounit.
In the medium term, I'll try to get a consensus on the best metrics to create a function like ScaleOnRequestType for NLBs.
Upvotes: 1