AWS Step Functions Tutorial

---

1. Creating a Simple Step Function

Example 1: Define a Basic Step Function

{
  "StartAt": "FirstStep",
  "States": {
    "FirstStep": {
      "Type": "Pass",
      "Next": "SecondStep"
    },
    "SecondStep": {
      "Type": "Succeed"
    }
  }
}

Explanation

1. Defines the starting point of execution.

   • "StartAt": "FirstStep"

2. FirstStep acts as a pass-through state.

   • "Type": "Pass", "Next": "SecondStep"

3. SecondStep completes the execution successfully.

   • "Type": "Succeed"

4. Creates a basic sequential workflow.

   • "Next": "SecondStep"

---

Example 2: Invoke a Lambda Function

{
  "StartAt": "CallLambda",
  "States": {
    "CallLambda": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:us-east-1:123456789012:function:MyLambda",
      "End": true
    }
  }
}

Explanation

1. Defines a state that calls a Lambda function.

   • "Type": "Task", "Resource": "arn:aws:lambda:us-east-1:123456789012:function:MyLambda"

2. Uses an AWS Lambda ARN for invocation.

   • "Resource": "arn:aws:lambda:us-east-1:123456789012:function:MyLambda"

3. Ends the step function after execution.

   • "End": true"

4. Executes the function and stops the workflow.

   • "StartAt": "CallLambda"

---

Example 3: Choice State for Conditional Logic

{
  "StartAt": "CheckValue",
  "States": {
    "CheckValue": {
      "Type": "Choice",
      "Choices": [
        {
          "Variable": "$.value",
          "NumericGreaterThan": 10,
          "Next": "Greater"
        }
      ],
      "Default": "LessThanOrEqual"
    },
    "Greater": { "Type": "Succeed" },
    "LessThanOrEqual": { "Type": "Succeed" }
  }
}

Explanation

1. Checks if the input value is greater than 10.

   • "Variable": "$.value", "NumericGreaterThan": 10

2. Routes execution to the ‘Greater’ state if true.

   • "Next": "Greater"

3. Defaults to ‘LessThanOrEqual’ if condition fails.

   • "Default": "LessThanOrEqual"

4. Enables branching workflows in Step Functions.

   • "Type": "Choice"

---

Example 4: Parallel Execution

{
  "StartAt": "ParallelExecution",
  "States": {
    "ParallelExecution": {
      "Type": "Parallel",
      "Branches": [
        {
          "StartAt": "Task1",
          "States": { "Task1": { "Type": "Succeed" } }
        },
        {
          "StartAt": "Task2",
          "States": { "Task2": { "Type": "Succeed" } }
        }
      ],
      "End": true
    }
  }
}

Explanation

1. Executes multiple branches simultaneously.

   • "Type": "Parallel"

2. Defines two parallel branches, Task1 and Task2.

   • "Branches": [{ "StartAt": "Task1" }, { "StartAt": "Task2" }]

3. Completes execution when both tasks finish.

   • "End": true"

4. Improves performance by running independent tasks together.

   • "States": { "Task1": { "Type": "Succeed" } }

---

2. Passing Data Between Steps

Example 1: Input Data for Step Function

{
  "StartAt": "ProcessData",
  "States": {
    "ProcessData": {
      "Type": "Pass",
      "Result": { "message": "Hello, World!" },
      "Next": "OutputData"
    },
    "OutputData": {
      "Type": "Succeed"
    }
  }
}

Explanation

1. Passes a static JSON object to the next step.

   • "Result": { "message": "Hello, World!" }

2. Ensures data flows from ‘ProcessData’ to ‘OutputData’.

   • "Next": "OutputData"

3. Useful for debugging and testing workflows.

   • "Type": "Pass"

4. Completes the workflow when OutputData executes.

   • "Type": "Succeed"

---

Example 2: Use InputPath to Filter Incoming Data

{
  "StartAt": "FilterData",
  "States": {
    "FilterData": {
      "Type": "Pass",
      "InputPath": "$.details",
      "Next": "OutputFilteredData"
    },
    "OutputFilteredData": { "Type": "Succeed" }
  }
}

Explanation

1. Extracts only the ‘details’ field from input.

   • "InputPath": "$.details"

2. Removes unnecessary input fields.

   • "Type": "Pass"

3. Sends the filtered data to the next state.

   • "Next": "OutputFilteredData"

4. Ensures clean, optimized data processing.

   • "Type": "Succeed"

---

Example 3: Use ResultPath to Modify Output

{
  "StartAt": "TransformData",
  "States": {
    "TransformData": {
      "Type": "Pass",
      "Result": { "status": "processed" },
      "ResultPath": "$.processingInfo",
      "Next": "Done"
    },
    "Done": { "Type": "Succeed" }
  }
}

Explanation

1. Adds new data without modifying original input.

   • "ResultPath": "$.processingInfo"

2. Stores processed status separately.

   • "Result": { "status": "processed" }

3. Combines existing and new data in workflow.

   • "Next": "Done"

4. Useful for transforming input data.

   • "Type": "Pass"

---

Example 4: Using OutputPath

{
  "StartAt": "TrimData",
  "States": {
    "TrimData": {
      "Type": "Pass",
      "OutputPath": "$.trimmed",
      "Next": "Finish"
    },
    "Finish": { "Type": "Succeed" }
  }
}

Explanation

1. Reduces the output to only the ‘trimmed’ field.

   • "OutputPath": "$.trimmed"

2. Removes unnecessary data from the workflow output.

   • "Type": "Pass"

3. Passes only the needed data to Finish state.

   • "Next": "Finish"

4. Helps optimize Step Function execution costs.

   • "Type": "Succeed"

---

Next Steps

• You can integrate Step Functions with Lambda, DynamoDB, SNS, and SQS.

• More advanced scenarios, are: error handling, retries, or state machine triggers, etc.