codedeploy-ecs-handel-extension

This repository contains a Handel extension which creates a CodeDeploy application and a multi-service
ECS cluster. The CodeDeploy application deploys to the ECS cluster using the blue/green deployment type.

Motivation

Configuring a multi-service ECS cluster requires quite a bit of configuration
(load balancer, target groups, security groups, routing, etc.).
This extension configures all of this for you. This extension also takes care of setting up a blue/green
deployment in CodeDeploy.

To support CodeDeploy blue/green deployments, this extension will generate a
lambda which can be used by CodeDeploy to trigger runscope tests. This lambda looks
in the AWS parameter store for the following variables
..runscope-trigger-url and
..runscope-access-token.

Another lambda is also generated which can be run manually to create a new deployment to your ECS cluster.
The name of this lambda is -create-deployment.

Usage

To use this extension, add it to the extensions section of your Handel file,
and then add the ecs service to your environment:

``yaml version: 1

name: codedeploy-ecs-example

extensions: # This tells Handel to import this extension codedeploy: codedeploy-ecs-handel-extension environments: dev: ecsCluster: type: 'codedeploy::ecs' # You must use the :: syntax here cluster: # Optional. key_name: example-key # Optional. The key used to ssh into this service instance_type: t2.small # Optional. default is 't2.micro' services: # Required. - name: service1 # Required. launch_type: EC2 # Optional. Default is 'EC2' auto_scaling: # Required. min_tasks: 2 # Required. This extension assumes that the desired task count = min_tasks max_tasks: 4 # Required. scaling_policies: # Optional. Configure scaling policies - type: up adjustment: value: 1 alarm: metric_name: CPUUtilization comparison_operator: GreaterThanThreshold threshold: 70 - type: down adjustment: value: 1 alarm: metric_name: CPUUtilization comparison_operator: GreaterThanThreshold threshold: 30 task_definition: # Required. name: exampleTask # Required. network_mode: bridge # Optional. Default is 'bridge' container: # Required. name: exampleContainer # Required image_name: brogerm/handel-multi-service-ecs-example # Optional. The Docker image that should be executed as the task max_mb: 2048 # Optional. Default: 128. The amount of memory in MB to allocate to the task cpu_units: 1024 # Optional. Default: 100. The amount of CPU units to allocate to the task. 1 vCPU = 1024 port_mappings: # Optional. - 8081 routing: # Optional. This will configure a load balancer listener rule to route traffic between services base_path: '/basepath1/*' health_check_path: '/xhealth' environment_variables: # Optional. REGION: us-west-2 codedeploy: termination_wait_time: 30 # Optional. Specify how long (in minutes) to wait before terminating old instances. Default is 60 time_before_auto_stop: 30 # Optional. Specify how long (in minutes) to wait before the deployment will be stopped automatically. # For a deployment to succeed, the user must manually choose to continue after the test endpoint is configured. # Default is 60 runscope_tests: # Optional. Keep in mind that AWS Lambda has a 15 minute maximum timeout, meaning that your max_allow_retries multiplied # by your retry_delay should not be more than 15 minutes and should also allow time for the tests to run. max_allowable_retries: 3 # Required. The number of maximum times to retry the runscope tests if they fail. Default is 0 retry_delay: 60 # Required. The amount of time, in seconds, to wait between retries. Default is 0 # See https://handel.readthedocs.io/en/latest/supported-services/ecs.html for configuring load balancing, logging, and tagging`

`Image Names`


This extension only supports images stored in your AWS account's EC2 Container Registry (ECR)

Reference an image as follows:`

`The piece will be replaced with your account's long ECR repository name`


/codedeploy-ecs-example

Notes

Since CloudFormation does not yet support creating CodeDeploy Blue/Green ECS deployments, the
ECS services and the CodeDeploy application were created using the aws-sdk. This means
that simply deleting the CloudFormation stack will not delete these resources. These resources
must be deleted manually.

Motivation

Another lambda is also generated which can be run manually to create a new deployment to your ECS cluster.
The name of this lambda is -create-deployment.

Usage

extensions

ecs

yaml
version: 1

name: codedeploy-ecs-example

extensions: # This tells Handel to import this extension
  codedeploy: codedeploy-ecs-handel-extension
    
environments:
  dev:
    ecsCluster:
      type: 'codedeploy::ecs' # You must use the :: syntax here
      cluster: # Optional.
        key_name: example-key # Optional. The key used to ssh into this service
        instance_type: t2.small # Optional. default is 't2.micro'
      services: # Required. 
      - name: service1 # Required.
        launch_type: EC2 # Optional. Default is 'EC2'
        auto_scaling: # Required.
          min_tasks: 2 # Required. This extension assumes that the desired task count = min_tasks
          max_tasks: 4 # Required.
          scaling_policies: # Optional. Configure scaling policies
          - type: up
            adjustment:
              value: 1
            alarm:
              metric_name: CPUUtilization
              comparison_operator: GreaterThanThreshold
              threshold: 70
          - type: down
            adjustment:
              value: 1
            alarm:
              metric_name: CPUUtilization
              comparison_operator: GreaterThanThreshold
              threshold: 30
        task_definition: # Required.
          name: exampleTask # Required.
          network_mode: bridge # Optional. Default is 'bridge'
          container: # Required.
            name: exampleContainer # Required
            image_name: brogerm/handel-multi-service-ecs-example # Optional. The Docker image that should be executed as the task
            max_mb: 2048 # Optional. Default: 128. The amount of memory in MB to allocate to the task
            cpu_units: 1024 # Optional. Default: 100. The amount of CPU units to allocate to the task. 1 vCPU = 1024
            port_mappings: # Optional. 
              - 8081
            routing: # Optional. This will configure a load balancer listener rule to route traffic between services
              base_path: '/basepath1/*'
              health_check_path: '/xhealth'
            environment_variables: # Optional.
              REGION: us-west-2
      codedeploy:
        termination_wait_time: 30 # Optional. Specify how long (in minutes) to wait before terminating old instances. Default is 60
        time_before_auto_stop: 30 # Optional. Specify how long (in minutes) to wait before the deployment will be stopped automatically. 
                                  # For a deployment to succeed, the user must manually choose to continue after the test endpoint is configured.
                                  # Default is 60
      runscope_tests: # Optional. Keep in mind that AWS Lambda has a 15 minute maximum timeout, meaning that your max_allow_retries multiplied
                      # by your retry_delay should not be more than 15 minutes and should also allow time for the tests to run.
        max_allowable_retries: 3 # Required. The number of maximum times to retry the runscope tests if they fail. Default is 0
        retry_delay: 60 # Required. The amount of time, in seconds, to wait between retries. Default is 0                       
      # See https://handel.readthedocs.io/en/latest/supported-services/ecs.html for configuring load balancing, logging, and tagging