autoscaling.yaml

SenzaInfo:
  StackName: hello-world
  StackVersion: one

SenzaComponents:
  - Configuration:
      Type: Senza::StupsAutoConfiguration # auto-detect network setup

  - AppServer:
      Type: Senza::TaupageAutoScalingGroup
      InstanceType: t2.micro
      IamRoles:
        - app-hello-world
      TaupageConfig:
        ports:
          8080: 8080
        application_version: "one"
        runtime: Docker
        source: "stups/hello-world:one"
      AssociatePublicIpAddress: true
      AutoScaling:
        # also possible: CPU, NetworkOut
        MetricType: NetworkIn
        # at least two instances
        Minimum: 2
        # at most sixteen instances
        Maximum: 16
        # successful startup requirement for the ASG:
        # if there are not at least two healthy instances
        # within 5 minutes timeout, the ASG creation is
        # considered failed.
        # default instance count is 1
        # default timeout is 15 minutes
        SuccessRequires: "2 within 5m"
        # scale two instances up/down at a time (if not overwritten by scale up/down specififc value)
        ScalingAdjustment: 2
        # scale three instances up at a time (if set, it overwrites the ScalingAdjustment value for scaling up)
        ScaleUpAdjustment: 3
        # scale one instance down at a time (if set, it overwrites the ScalingAdjustment value for scaling down)
        ScaleDownAdjustment: 1
        # after scaling, suspend further scaling activities
        # for 30 minutes (1800 seconds)
        # Can be overwritten by scale up/down specific values
        Cooldown: 1800
        # If set it overwrites the Cooldown value for scaling up
        ScaleUpCooldown: 300
        # If set it overwrites the Cooldown value for scaling down
        ScaleDownCooldown: 3600
        # time over which the threshold is measured in seconds
        Period: 300
        # how many periods the threshold must be hit before scaling
        # this will not scale earlier than 10 mins (2 * 300s)
        EvaluationPeriods: 2
        # will scale up once >= 20 GB inbound traffic
        # are hit over 10 mins on average (evaluation periods * period)
        ScaleUpThreshold: "20 GB"
        # will scale down once <= 1 GB inbound traffic
        # is hit over 10 mins on average (evaluation periods * period)
        ScaleDownThreshold: "1 GB"