Skip to main content

​
Prerequisites

Ensure you define the LatestAmiId parameter before deploying the stack. For example:
Parameters:
  LatestAmiId:
    Type: AWS::SSM::Parameter::Value<AWS::EC2::Image::Id>
    Default: /aws/service/ami-amazon-linux-latest/amzn2-ami-hvm-x86_64-gp2

​
CloudFormation Template Overview

This CloudFormation template provisions the following AWS resources:
ResourceDescriptionCloudFormation Type
Security GroupSSH access for EC2 instancesAWS::EC2::SecurityGroup
Launch ConfigurationLaunches t2.micro instances with latest AMIAWS::AutoScaling::LaunchConfiguration
Auto Scaling Group (ASG)Scales between 2 and 20 instances, 1-minute metricsAWS::AutoScaling::AutoScalingGroup
Scale-Up PolicyIncreases ASG capacity by 1 instanceAWS::AutoScaling::ScalingPolicy
CloudWatch Alarm: HighCPUTriggers when CPU > 80%AWS::CloudWatch::Alarm
CloudWatch Alarm: VeryHighCPUTriggers when CPU > 90%AWS::CloudWatch::Alarm
CloudWatch Alarm: HighASGSizeTriggers when ASG size > 10AWS::CloudWatch::Alarm

​
Core Resources

Resources:
  MyEC2SecurityGroup:
    Type: 'AWS::EC2::SecurityGroup'
    Properties:
      GroupDescription: EC2 Security Group for the Auto Scaling Group instances
      SecurityGroupIngress:
        - IpProtocol: tcp
          FromPort: '22'
          ToPort: '22'
          CidrIp: 0.0.0.0/0

  LaunchConfiguration:
    Type: 'AWS::AutoScaling::LaunchConfiguration'
    Properties:
      ImageId: !Ref LatestAmiId
      InstanceType: t2.micro
      SecurityGroups:
        - Ref: MyEC2SecurityGroup

  AutoScalingGroup:
    Type: 'AWS::AutoScaling::AutoScalingGroup'
    Properties:
      AvailabilityZones: Fn::GetAZs: ''
      LaunchConfigurationName: Ref: LaunchConfiguration
      MinSize: '2'
      MaxSize: '20'
      DesiredCapacity: '2'
      MetricsCollection:
        - Granularity: 1Minute

  ScaleUpPolicy:
    Type: 'AWS::AutoScaling::ScalingPolicy'
    Properties:
      AutoScalingGroupName: Ref: AutoScalingGroup
      AdjustmentType: ChangeInCapacity
      ScalingAdjustment: '1'
      Cooldown: '300'

​
CloudWatch Alarms Definition

Alarm NameMetricThresholdPeriodNamespace
HighCPUUsageCPUUtilization80%300sAWS/EC2
VeryHighCPUUsageCPUUtilization90%300sAWS/EC2
HighASGSizeGroupTotalInstances10300sAWS/AutoScaling
  HighCPUAlarm:
    Type: 'AWS::CloudWatch::Alarm'
    Properties:
      AlarmName: HighCPUUsage
      AlarmDescription: Alarm when CPU exceeds 80 percent
      Namespace: AWS/EC2
      MetricName: CPUUtilization
      Statistic: Average
      Period: '300'
      EvaluationPeriods: '1'
      Threshold: '80'
      ComparisonOperator: GreaterThanThreshold
      Dimensions:
        - Name: AutoScalingGroupName
          Value: Ref: AutoScalingGroup

  VeryHighCPUAlarm:
    Type: 'AWS::CloudWatch::Alarm'
    Properties:
      AlarmName: VeryHighCPUUsage
      AlarmDescription: Alarm when CPU exceeds 90 percent
      Namespace: AWS/EC2
      MetricName: CPUUtilization
      Statistic: Average
      Period: '300'
      EvaluationPeriods: '1'
      Threshold: '90'
      ComparisonOperator: GreaterThanThreshold
      Dimensions:
        - Name: AutoScalingGroupName
          Value: Ref: AutoScalingGroup

  HighASGSizeAlarm:
    Type: 'AWS::CloudWatch::Alarm'
    Properties:
      AlarmName: HighASGSize
      AlarmDescription: Alarm when ASG size exceeds 10
      Namespace: AWS/AutoScaling
      MetricName: GroupTotalInstances
      Statistic: Average
      Period: '300'
      EvaluationPeriods: '1'
      Threshold: '10'
      ComparisonOperator: GreaterThanThreshold
      Dimensions:
        - Name: AutoScalingGroupName
          Value: Ref: AutoScalingGroup

​
Outputs

Outputs:
  AutoScalingGroupName:
    Description: The name of the Auto Scaling Group
    Value: Ref: AutoScalingGroup
Although the ASG can scale up to 20 instances, our threshold is set at >10 for demonstration.

​
Deploy the CloudFormation Stack

  1. Open the AWS CloudFormation Console.
  2. Choose Create stack > Upload a template file, and select your composite_alarm.yaml.
The image shows an AWS CloudFormation interface for creating a stack, with a file selection window open to choose a YAML file named "composite_alarm.yaml" from the Downloads folder.
  1. Click Next, set Stack name to CompositeAlarmInfra, and keep default parameters.
The image shows an AWS CloudFormation interface where a user is specifying stack details, including a stack name and parameters for creating a stack.
  1. Submit and wait ~4–5 minutes until status is CREATE_COMPLETE.
The image shows an AWS CloudFormation console with a stack named "composite-alarm-infra" and a list of events indicating the creation status of various resources.
  1. On Outputs, copy the AutoScalingGroupName.
The image shows an AWS CloudFormation console with a stack named "composite-alarm-infra" that has a status of "CREATE_COMPLETE." The Outputs tab displays an AutoScalingGroupName with its corresponding value.

​
Verify CloudWatch Alarms

Navigate to CloudWatch > Alarms. You should see:
  • HighASGSize
  • HighCPUUsage
  • VeryHighCPUUsage
Initial states will be OK or Insufficient Data.
The image shows an AWS CloudWatch dashboard displaying alarms for EC2 and AutoScaling services, with recent alarms indicating high CPU usage and auto-scaling group size.
The image shows an AWS CloudWatch Alarms dashboard with three alarms listed, indicating their status and conditions. The alarms are named "HighASGSize," "HighCPUUsage," and "VeryHighCPUUsage," with varying states and conditions.

​
Create the Composite Alarm

  1. Select the three child alarms.
  2. Go to Actions > Create composite alarm.
  3. Define the rule: 
    ALARM(HighCPUUsage) OR ALARM(VeryHighCPUUsage) OR ALARM(HighASGSize)
  4. Configure notifications (e.g., an SNS topic) and an optional suppressor alarm.
The image shows an AWS CloudWatch console screen where a user is configuring alarm notifications and actions, including selecting an SNS topic and managing alarm actions.
  1. Enter Name and Description (Markdown supported).
The image shows an AWS CloudWatch interface for creating a composite alarm, specifically the "Add name and description" step. It includes fields for entering the alarm name and an optional description with markdown formatting guidelines.
After creation, you’ll see four alarms, all in OK state.
The image shows an AWS CloudWatch dashboard with a list of alarms, all in an "OK" state, detailing conditions like CPU utilization and actions enabled.

​
Test the Composite Alarm

To trigger the alarm:
  1. In the EC2 Console, open your ASG > Details > Edit.
  2. Set Min and Desired capacity to 12, then save.
The image shows an AWS EC2 Auto Scaling group management console with two instances listed as "InService" and "Healthy." The interface includes options for managing lifecycle hooks and warm pools.
After one metric period, HighASGSize will fire and the composite alarm will trigger.
The image shows an AWS CloudWatch dashboard displaying a composite alarm with its status as "OK" and a list of child alarms, all of which are also in the "OK" state. The interface includes options for viewing details, tags, alarm rules, actions, and history.
The image shows an AWS CloudWatch dashboard with alarms, including a composite alarm that is currently in an alarm state. The timeline and child alarms are displayed, indicating their status and last state changes.
Check your email for the SNS notification.

​
Cleanup

  1. Delete the CloudFormation stack.
The image shows an AWS CloudFormation console with details of a stack named "composite-alarm-infra" that is in the process of being deleted. The status is "DELETE_IN_PROGRESS" and it includes information like stack ID, description, and creation time.
  1. In CloudWatch, select the composite alarm and choose Actions > Delete.
Deleting the composite alarm must be done manually—it is not removed by the CloudFormation stack.
This completes the guide to building and testing an AWS composite alarm using CloudFormation, Auto Scaling, and CloudWatch.

Watch Video

Practice Lab