Choosing Step Functions workflow type - AWS Step Functions

Choosing Step Functions workflow type

When you create a state machine, you select a Type of either Standard or Express. The default Type for state machines is Standard. A state machine whose Type is Standard is called a Standard workflow and a state machine whose Type is Express is called an Express workflow.

For both Standard and Express workflows, you define your state machine using the Using Amazon States Language to define Step Functions workflows. Your state machine executions will behave differently depending on the Type that you select.

Important

The Type you choose can't be changed after you create the state machine.

Note

If you define your state machines outside the Step Functions' console, such as in an editor of your choice, you must save your state machine definitions with the extension .asl.json.

Standard Workflows are ideal for long-running (up to one year), durable, and auditable workflows. You can retrieve the full execution history using the Step Functions API for up to 90 days after your execution completes. Standard Workflows follow an exactly-once model, where your tasks and states are never run more than once, unless you have specified Retry behavior in ASL. This makes Standard Workflows suited to orchestrating non-idempotent actions, such as starting an Amazon EMR cluster or processing payments. Standard Workflow executions are billed according to the number of state transitions processed.

Express Workflows are ideal for high-volume, event-processing workloads such as IoT data ingestion, streaming data processing and transformation, and mobile application backends. They can run for up to five minutes. Express Workflows employ an at-least-once model, where an execution could potentially run more than once. This makes Express Workflows ideal for orchestrating idempotent actions such as transforming input data and storing by way of a PUT action in Amazon DynamoDB. Express Workflow executions are billed by the number of executions, the duration of execution, and the memory consumed while the execution ran.

Standard and Express Workflows can automatically start in response to events such as HTTP requests from Amazon API Gateway (fully-managed APIs at scale), IoT Rules and over 140 other event sources in Amazon EventBridge.

Tip

To deploy an example of an Express workflow to your AWS account, see Module 7 - API Gateway, Parallel State, Express workflows of The AWS Step Functions Workshop.

For information about the console experience for Standard and Express Workflow executions, see Viewing workflow execution details in Step Functions.

Standard vs Express Workflows
Standard Workflows Express Workflows: Synchronous and Asynchronous
Maximum duration One year Five minutes
Supported execution start rate

For information about quotas related to supported execution start rate, see Quotas related to API action throttling.

For information about quotas related to supported execution start rate, see Quotas related to API action throttling.

Supported state transition rate

For information about quotas related to supported state transition rate, see Quotas related to state throttling.

No limit
Pricing Priced by number of state transitions. A state transition is counted each time a step in your execution is completed. Priced by the number of executions you run, their duration, and memory consumption.
Execution history

Executions can be listed and described with Step Functions APIs. Executions can be visually debugged through the console. They can also be inspected in CloudWatch Logs by enabling logging on your state machine.

For more information about debugging Standard Workflow executions in the console, see Viewing workflow execution details in Step Functions and Viewing and debugging executions.

Unlimited execution history, that is, as many execution history entries are maintained as you can generate within a 5-minute period.

Executions can be inspected in CloudWatch Logs or the Step Functions console by enabling logging on your state machine.

For more information about debugging Express Workflow executions in the console, see Viewing workflow execution details in Step Functions and Viewing and debugging executions.

Execution semantics Exactly-once workflow execution.

Asynchronous Express Workflows: At-least-once workflow execution.

Synchronous Express Workflows: At-most-once workflow execution.

Service integrations Supports all service integrations and patterns. Supports all service integrations.
Note

Express Workflows do not support Job-run (.sync) or Callback (.waitForTaskToken) service integration patterns.

Step Functions activities Supports Step Functions activities. Doesn't support Step Functions activities.

Synchronous and Asynchronous Express Workflows in Step Functions

There are two types of Express Workflows that you can choose: Asynchronous Express Workflows and Synchronous Express Workflows.

  • Asynchronous Express Workflows return confirmation that the workflow was started, but don't wait for the workflow to complete. To get the result, you must poll the service's CloudWatch Logs. You can use Asynchronous Express Workflows when you don't require immediate response output, such as messaging services or data processing that other services don't depend on. You can start Asynchronous Express Workflows in response to an event, by a nested workflow in Step Functions, or by using the StartExecution API call.

  • Synchronous Express Workflows start a workflow, wait until it completes, and then return the result. Synchronous Express Workflows can be used to orchestrate microservices. With Synchronous Express Workflows, you can develop applications without the need to develop additional code to handle errors, retries, or run parallel tasks. You can run Synchronous Express Workflows invoked from Amazon API Gateway, AWS Lambda, or by using the StartSyncExecution API call.

    Note

    If you run Step Functions Express Workflows synchronously from the console, the StartSyncExecution request elapses after 60 seconds. To run the Express Workflows synchronously for a duration of up to five minutes, make the StartSyncExecution request using the AWS SDK or AWS Command Line Interface (AWS CLI) instead of the Step Functions console.

    Synchronous Express execution API calls don't contribute to existing account capacity limits. Step Functions provides capacity on demand and automatically scales with sustained workload. Surges in workload may be throttled until capacity is available.

Execution guarantees in Step Functions workflows

Standard Workflows Asynchronous Express Workflows Synchronous Express Workflows
Exactly-once workflow execution At-least-once workflow execution At-most-once workflow execution
Execution state internally persists between state transitions. Execution state doesn't persist between state transitions. Execution state doesn't persist between state transitions.
Automatically returns an idempotent response on starting an execution with the same name as a currently-running workflow. The new workflow doesn't start and an exception is thrown once the currently-running workflow is complete. Idempotency is not automatically managed. Starting multiple workflows with the same name results in concurrent executions. Can result in loss of internal workflow state if state machine logic is not idempotent. Idempotency is not automatically managed. Step Functions waits once an execution starts and returns the state machine's result on completion. Workflows don't restart if an exception occurs.

Execution history data removed after 90 days. Workflow names can be reused after removal of out-of-date execution data.

To meet compliance, organizational, or regulatory requirements, you can reduce the execution history retention period to 30 days by sending a quota request. To do this, use the AWS Support Center Console and create a new case.

Execution history is not captured by Step Functions. Logging must be enabled through Amazon CloudWatch Logs. Execution history is not captured by Step Functions. Logging must be enabled through Amazon CloudWatch Logs.