You are the DevOps engineer in a company assigned to support a development team interested in having a CI/CD pipeline in AWS CodePipeline to deal with several complex cases, handle asynchronous workflow-driven actions, and keep centralized logging for the workflow execution history to CloudWatch Logs. What is the most robust approach to follow?

Answer options:

A.Create an AWS Lambda function to handle workflow-driven actions. Configure a CodePipeline deploy action to trigger the same AWS Lambda function. To deal with centralized logging, an AWS Step Function can be configured to log workflow execution history to CloudWatch Logs.
B.Create an AWS Step Functions state machine to handle workflow-driven actions via specific, lean single purpose AWS Lambda functions. Configure a CodePipeline deploy action to trigger an AWS Lambda function that decouples the Step Functions state machine and the pipeline itself. To deal with centralized logging, it is possible while creating or updating a state machine to configure logging to Amazon CloudWatch Logs.
C.Create an AWS Step Functions state machine to handle workflow-driven actions via specific, lean single purpose AWS Lambda functions. Configure a CodePipeline deploy action to trigger an AWS Lambda function that decouples the Step Functions state machine and the pipeline itself. To deal with centralized logging, an Amazon S3 bucket can be defined to store all logs since there is no integration between AWS Step Functions and Amazon CloudWatch Logs.
D.Create an AWS Lambda function to handle workflow-driven actions. Configure a CodePipeline deploy action to trigger the same AWS Lambda function. To deal with centralized logging, an Amazon S3 bucket can be defined to store all logs from Amazon CloudWatch Logs.

Correct Answer: B
The power of AWS Step Functions is orchestrating workflow-driven actions and joining other services based on logic in the workflow. A robust way to decouple AWS Step Functions and AWS CodePipeline are via an AWS Lambda function triggered by AWS CodePipeline and acting in the interim between workflow actions and the pipeline itself. As of February 2020, AWS Step Functions support AWS CloudWatch Logs to log workflow execution history, which makes it possible to centralize logging to this service as required in the problem description.
Incorrect Answers:
Options A, D are incorrect because AWS Lambda functions are not the most robust approach to handle workflow-driven actions in this scenario.
Option C is incorrect because the problem description states using AWS CloudWatch Logs as a centralized logging service instead of Amazon S3. AWS CloudWatch Logs is supported as of February 2020 by AWS Step Functions to log workflow execution history.
References:
https://aws.amazon.com/blogs/devops/using-aws-step-functions-state-machines-to-handle-workflow-driven-aws-codepipeline-actions/
https://aws.amazon.com/about-aws/whats-new/2020/02/aws-step-functions-supports-cloudwatch-logs-standard-workflows/