An enterprise is designing a deployment pipeline for a critical containerized microservice… — AWS Certified Solutions Architect - Professional

An enterprise is designing a deployment pipeline for a critical containerized microservice running on Amazon ECS with AWS Fargate behind an Application Load Balancer (ALB). The infrastructure and deployment pipelines must be managed as infrastructure as code (IaC) using AWS CloudFormation. The enterprise has established the following requirements for deployment strategy design:
- Deployments must be fully automated and must not cause any application downtime.
- A new version of the microservice must be tested with a subset of live traffic, routing exactly $10\%$ of the production traffic to the new version for a duration of $15$ minutes before shifting the remaining $90\%$ of traffic.
- The deployment must automatically roll back immediately if the new version displays any degradation in performance, specifically measured by an increase in $5\text{xx}$ response codes.

Which two configuration steps should the solutions architect combine to meet these requirements?

Configure the AWS CloudFormation template to use the `AWS::CodeDeploy::BlueGreen` transform, defining an Amazon ECS service that references an AWS CodeDeploy deployment group configured with the `CodeDeployDefault.ECSCanary10Percent15Minutes` deployment configuration.Answer
Create an Amazon CloudWatch alarm to monitor the `HTTPCode_Target_5XX_Count` metric of the replacement target group on the Application Load Balancer, and configure the CodeDeploy deployment group to perform an automatic rollback when this alarm enters the `ALARM` state.Answer
C
Enable the Amazon ECS deployment circuit breaker on the service definition in the CloudFormation template, and configure the `rollback` parameter to true to automatically roll back the service if container health checks or $5\text{xx}$ error counts exceed the threshold.
D
Configure an `UpdatePolicy` with `AutoScalingRollingUpdate` on the Amazon ECS service resource in the CloudFormation template to shift traffic progressively in increments of $10\%$ every $1.5$ minutes.
E
Create two Route $53$ weighted alias records pointing to separate blue and green Application Load Balancers, and implement an AWS Lambda function triggered by a cron schedule to increment the weight of the green record by $10\%$ every $1.5$ minutes.

Answer

Configure the AWS CloudFormation template to use the `AWS::CodeDeploy::BlueGreen` transform with the `CodeDeployDefault.ECSCanary10Percent15Minutes` configuration, and create an Amazon CloudWatch alarm on the `HTTPCode_Target_5XX_Count` metric of the replacement target group to trigger automatic rollback in CodeDeploy.

Implementing a blue/green deployment for Amazon ECS with AWS CloudFormation requires using the `AWS::CodeDeploy::BlueGreen` hook. This integrates CloudFormation with AWS CodeDeploy, allowing the use of predefined deployment configurations like `CodeDeployDefault.ECSCanary10Percent15Minutes` to route

10\%

of the traffic to the replacement task set for

15

minutes before shifting the remaining

90\%

. To automate rollbacks based on application-level issues, a CloudWatch alarm must be created to track

5\text{xx}

errors on the replacement (green) target group. Linking this alarm to the CodeDeploy deployment group's automatic rollback configuration ensures that CodeDeploy will immediately revert the traffic shift if the new version exhibits failure rates during the test phase.

Step-by-Step Solution

Define the blue/green deployment mechanism for the ECS Fargate service in CloudFormation.

Use the `AWS::CodeDeploy::BlueGreen` transform to delegate deployment management to AWS CodeDeploy.

CloudFormation alone does not natively support canary deployments for ECS; it requires integration with CodeDeploy to orchestrate traffic shifting.

Select the appropriate traffic routing configuration.

Specify the `CodeDeployDefault.ECSCanary10Percent15Minutes` deployment configuration in the CodeDeploy deployment group.

This configuration shifts

10\%

of traffic to the green task set for

15

minutes, then shifts the remaining

90\%

if no errors are detected, meeting the traffic routing requirement.

Configure the rollback trigger based on error rates.

Create a CloudWatch alarm monitoring `HTTPCode_Target_5XX_Count` on the replacement target group and configure the CodeDeploy deployment group to roll back when the alarm is in `ALARM` state.

This ensures that if the new version generates

5\text{xx}

errors during the

15

-minute test period, CodeDeploy will automatically roll back to the old version immediately.

Key Concept

Integration of AWS CloudFormation with AWS CodeDeploy for automated, alarm-backed container canary deployments.

Estimated Time:3m 0s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic