AWS Step Functions
AWS serverless orchestration service that coordinates multiple services into visual workflows, with built-in error handling, retries, and parallel execution.
seed#aws#step-functions#orchestration#serverless#workflow#state-machine
What it is
AWS Step Functions is an orchestration service that coordinates multiple AWS services into serverless workflows. It defines flows as state machines with steps, conditions, parallelism, and error handling — all visual and declarative.
Workflow types
| Type | Duration | Cost | Use |
|---|---|---|---|
| Standard | Up to 1 year | Per transition | Long, durable workflows |
| Express | Up to 5 min | Per execution/duration | High volume, low latency |
Available states
| State | Function | Example |
|---|---|---|
| Task | Execute work | Lambda, ECS, API call |
| Choice | Conditional branching | Validate input, route by type |
| Parallel | Parallel branch execution | Process data + send notification |
| Map | Iteration over arrays | Process each item in an order |
| Wait | Time-based pause | Wait 24h before retry |
| Pass | Data transformation | Reformat payload between steps |
| Succeed/Fail | Explicit termination | Mark workflow as completed or failed |
Common patterns
- Saga pattern: distributed transactions with compensation
- Human approval: pause workflow awaiting approval
- ETL pipelines: data processing in steps
- AI/ML pipelines: orchestrate training and inference
Why it matters
Step Functions turns complex workflows into visual, auditable state machines. Instead of orchestrating Lambda with custom code, you define the flow declaratively with built-in error handling, retries, and parallelism.
References
- Step Functions Documentation — Official documentation.
- Serverless Land - Step Functions — Patterns.
- Step Functions Best Practices — AWS, 2024. Best practices for state machines.