Skip to content

Latest commit

 

History

History
66 lines (44 loc) · 4.25 KB

task-1.md

File metadata and controls

66 lines (44 loc) · 4.25 KB

Task 1 - Start the workflow

Where do we start?

Checkout onto branch task/1 unless already there:

    git checkout task/1

In the serverless.yml file there's an inputBucket defined in serverless.yml.

There's also a scan-file lambda function registered:

  • see functions/scan-file/function.yml
  • see how it subscribes to a ObjectCreated event from S3.

This means that the lambda function gets fired every time a file is uploaded into the S3 bucket.

Try yourself what the lambda's input arguments are when it gets invoked this way:

  1. Add a log entry in functions/scan-file/handler.ts to log event argument of the handler method.
  2. Deploy the Serverless stack: npx sls deploy --aws-profile=tsh-workshops.
  3. In the AWS S3 Console choose the inputBucket which contains your username (eg. tsh-academy-serverless-workshops-input-bucket-student00 and upload any file from your computer ("Upload" -> "Add files" -> "Upload" button).
  4. In CloudWatch logs, see the logged output.

On top of that, serverless.yml also declares a step function workflow, named ScanCvWorkflow. Its definition is held in another file, workflows/scan-cv-workflow/workflow.asl.yml . There's only one step in it, which invokes an example Lambda function.

  • See how the Lambda function is referenced by its name, as it's defined in workflows/scan-cv-workflow/example-lambda/function.yml.
  • See how GetAtt Cloud Formation function is called to access the Lambda's ARN and provide it to the step definition.
  • Note that "Start" and "End" are not states per se, and you don't define them in the workflow.asl.yml file.
  • You can try out the workflow manually by clicking Start execution in the AWS Step Functions Console.

Description

What we want to achieve is that the workflow starts automatically as soon as a resume file is uploaded to the input S3 bucket.

We need you to add logic to scan-file lambda which will start the workflow execution:

  1. Open functions/scan-file/handler.ts .

  2. Create a new StepFunctions client, preferably outside the handle() method,

     const sf = new StepFunctions();

  3. In the handle method, invoke the ScanCvWorkflow step function, using the client's startExecution method:

     sf.startExecution({...});
  • as the execution's input, pass the uploaded file's path
  • name should be a unique identifier for each execution - you might use the filename followed by a uuid
  • stateMachineArn has already been provided as an env variable - see the config object
  • by the way, SDK methods have the utility method .promise() for use with async/await :)

  1. Redeploy the stack by running:

    npx sls deploy --aws-profile=tsh-workshops

Result

You may now upload one of the test files (see: ./test-data) into your input AWS bucket with the S3 Console.

Once you've uploaded a file, you should see:

  1. A workflow started in the Step Functions console
  2. Log entries from the scan-file lambda in Cloud Watch console (You need to find a LogGroup containing your username.)

Useful links