A DevOps engineer is building up a pipeline for a Java project which is under development. AWS CodePipeline service has been chosen as the CI/CD tool. For the service deployment phase, CloudFormation stacks are used. However, there are several situations to be considered. For example, (i) stack should be created if a stack does not exist, (ii) Change Set should be created and inspected if a stack already exists. The team lead prefers to implement a visualized state machine for the deployment and each state executes a Lambda function. What is the best way to achieve this requirement?

Answer options:

A.Establish several SQS queues to indicate running status. The Lambda function for each state gets the message from the queue and processes the deployment task accordingly. The final result is returned to CodePipeline by a dedicated Lambda.
B.Use a Python script in EC2 to record the running status and achieve the state machine feature. It calls different Lambda functions depending on the current status. The script returns the execution status back to CodePipeline.
C.Use a Lambda Function to interface with an AWS Step Functions which implements the workflow-driven state machines for CloudFormation stack deployment. The Lambda Function returns the execution status back to CodePipeline.
D.Use a shell script in EC2 to interface with the AWS MQ service to achieve the function of the state machine. Depending on the running status of the AWS MQ service, the script returns the execution result back to CodePipeline.

Correct Answer – C
A visualized state machine is required in this question. AWS Step Functions is the best tool to achieve this. About how to integrate with AWS CodePipeline for AWS Step Functions, check this document in
https://aws.amazon.com/blogs/devops/using-aws-step-functions-state-machines-to-handle-workflow-driven-aws-codepipeline-actions/.
Option A is incorrect: Because SQS is not ideal for implementing a state machine. It also cannot provide a native visualized status as Step Functions.
Option B is incorrect: Similar reason as
Option A.Option C is CORRECT: Because AWS Step Functions makes it easy for users to understand which tasks have been executed and why a state has been reached. It also can interact with Lambda and inform AWS CodePipeline of its status.
Option D is incorrect: Because AWS MQ service is a service for ActiveMQ that can set up and operate message brokers in AWS. It is not an ideal tool to achieve a work-flow driven state machine.