alexgt9 / training-step-functions

AWS Step functions

SFN What is that?

• Orquestrator
• State machine
• Choice
• Parallelisation

Before doing the exercise

I recommend to follow this AWS example before trying to do the next problem

Practice: Approving task state machine

The problem

Create an state machine and the needed lambdas to simulate this process:

• The code is created
• A branch is created
• Notify the reviewers (the message should contains the branch name)
• Create a manual lambda able to approve or reject at this point
• If reject, the state machine should finish in fail state
• If success, merge to master (based on the branch name)
• Create a release
• Notify Ops to deploy the release (the message should contains the release name created in the previous step)

* Actions can be simulated doing an http request to a site like https://webhook.site/
* For new data like code, branch and release, you can add it in the steps using statis data in the Payload
* The lambda should return at least the payload sent to them to be able to reuse the data in next steps

Solution:

Just check this if you really don't want to challenge yourself or after you have done the practice

1. Create the lambdas based on the code in /lambdas folder

The lambda approveStepWithToken.js needs access to act on stepFunctions, you can add the role AWSStepFunctionsFullAccess in this case

2. Create the State Machine using the definition of SolveIssueStepFunction.json
3. You can test it using the test events in /test-events folder
   • Start the State Machine with test-events/initialEventTest.json, change the path to point to your webhook.site path
   • Check the calls you received, the last one should have the token needed to continue the Notify Reviewers state
   • Use this token to call the lambda lambdas/approveStepWithToken.js using the test event test-events/approveTaskTest.json
   • You can approve or reject in the previous event to decide the next branch in the state machine
   • That's all! The State Machine should have finished and you will be able to visualize the whole workflow in the AWS console