Managed Orchestration: Step Functions and Friends

AWS Step Functions defines workflows as state machines (Amazon States Language) with built-in retries, catch/compensation, parallel branches, waits and direct integrations (Lambda, SQS, DynamoDB, ECS). Standard workflows run up to a year with exactly-once state transitions; Express workflows are cheaper and at-least-once for high volume. Similar: Google Workflows, Azure Logic Apps/Durable Functions. They trade code flexibility for zero infrastructure and visual tracing.

Workflow engines

{
  "StartAt": "ReserveStock",
  "States": {
    "ReserveStock": { "Type": "Task", "Resource": "arn:aws:states:::lambda:invoke",
      "Parameters": { "FunctionName": "reserve-stock", "Payload.$": "$" },
      "Retry": [{ "ErrorEquals": ["States.TaskFailed"], "IntervalSeconds": 2, "BackoffRate": 2, "MaxAttempts": 5, "JitterStrategy": "FULL" }],
      "Next": "ChargePayment" },
    "ChargePayment": { "Type": "Task", "Resource": "arn:aws:states:::lambda:invoke",
      "Parameters": { "FunctionName": "charge-payment", "Payload.$": "$" },
      "Catch": [{ "ErrorEquals": ["PaymentDeclined"], "Next": "ReleaseStock" }],
      "Next": "Ship" },
    "ReleaseStock": { "Type": "Task", "Resource": "arn:aws:states:::lambda:invoke",
      "Parameters": { "FunctionName": "release-stock", "Payload.$": "$" }, "Next": "Failed" },
    "Ship": { "Type": "Task", "Resource": "arn:aws:states:::sqs:sendMessage",
      "Parameters": { "QueueUrl": "https://sqs.eu-west-1.amazonaws.com/123/shipments", "MessageBody.$": "$" }, "End": true },
    "Failed": { "Type": "Fail" }
  }
}

Common pitfalls

• Complex business logic in giant JSON state machines.
• Express workflows for non-idempotent steps.
• Ignoring per-transition costs at high volume.

Key points

• State machines with retries/catch
• Standard vs Express semantics
• Code-first engines for complex logic

Further reading: AWS Step Functions