ECS Fargate CPU目标追踪自动扩缩容：低CPU利用率告警持续触发问题

算力阿峰

2026-7-13

问题

我运行着启用CPU目标追踪自动扩缩容的ECS Fargate服务，发现低CPU利用率的CloudWatch告警始终处于“告警中”状态。是否因为目标值设为70%，自动扩缩容告警会同时触发低于和高于70%的情况？我不希望该低CPU告警持续处于告警状态。

附CloudFormation配置代码：

#Policy type: Target Tracking scaling
#Disable scale-in: No

### ECS Service Target Tracking Auto Scaling Configuration
  ECSScalableTarget:
    Type: AWS::ApplicationAutoScaling::ScalableTarget
    Condition: SetTargetTrackingAutoScaling
    Properties:
      MaxCapacity: 1
      MinCapacity: 1
      ResourceId: "serviceName"
      RoleARN: !Sub 'arn:aws:iam::${AWS::AccountId}:role/aws-service-role/ecs.application-autoscaling.amazonaws.com/AWSServiceRoleForApplicationAutoScaling_ECSService' 
      ScalableDimension: 'ecs:service:DesiredCount'  ## 1 The initial desired count to start with before Service Auto Scaling begins adjustment.
      ServiceNamespace: ecs

  ServiceScalingPolicyCPU:
    Type: AWS::ApplicationAutoScaling::ScalingPolicy
    Condition: SetTargetTrackingAutoScalingCPU
    Properties:
      PolicyName: target-tracking-cpu
      PolicyType: TargetTrackingScaling
      ScalingTargetId: !Ref ECSScalableTarget
      TargetTrackingScalingPolicyConfiguration:
        PredefinedMetricSpecification:
          PredefinedMetricType: ECSServiceAverageCPUUtilization
        TargetValue: 70
        ScaleInCooldown: 180
        ScaleOutCooldown: 60

原因分析

你的判断正确：目标追踪扩缩容策略会自动生成两个CloudWatch告警——一个监控CPU利用率高于70%（触发扩容），另一个监控CPU利用率低于70%（触发缩容）。但你的ECS服务配置了MinCapacity: 1和MaxCapacity: 1，实例数被固定为1，无法执行缩容操作。当CPU利用率低于70%时，缩容告警触发，但因已达最小实例数无法执行对应操作，导致告警持续处于“告警中”状态。

解决方案

有两种可行处理方式：

调整扩缩容范围：若业务允许，将MinCapacity设为小于当前实例数的值（比如0），这样CPU利用率低于目标值时，服务可缩容到最小实例数，告警会自动恢复。修改ECSScalableTarget配置：

ECSScalableTarget:
  Type: AWS::ApplicationAutoScaling::ScalableTarget
  Condition: SetTargetTrackingAutoScaling
  Properties:
    MaxCapacity: 1  # 可根据业务需求调整
    MinCapacity: 0   # 修改为允许缩容到0
    ResourceId: "serviceName"
    RoleARN: !Sub 'arn:aws:iam::${AWS::AccountId}:role/aws-service-role/ecs.application-autoscaling.amazonaws.com/AWSServiceRoleForApplicationAutoScaling_ECSService' 
    ScalableDimension: 'ecs:service:DesiredCount'
    ServiceNamespace: ecs

禁用缩容逻辑：若业务需要保持实例数至少为1，不想开启缩容，可在目标追踪策略中添加DisableScaleIn: true，这样自动扩缩容只会创建扩容告警，不会生成低CPU的缩容告警。修改ServiceScalingPolicyCPU配置：

ServiceScalingPolicyCPU:
  Type: AWS::ApplicationAutoScaling::ScalingPolicy
  Condition: SetTargetTrackingAutoScalingCPU
  Properties:
    PolicyName: target-tracking-cpu
    PolicyType: TargetTrackingScaling
    ScalingTargetId: !Ref ECSScalableTarget
    TargetTrackingScalingPolicyConfiguration:
      PredefinedMetricSpecification:
        PredefinedMetricType: ECSServiceAverageCPUUtilization
      TargetValue: 70
      ScaleInCooldown: 180
      ScaleOutCooldown: 60
      DisableScaleIn: true  # 添加该参数禁用缩容

内容的提问来源于stack exchange，提问作者Sanjesh M