Application Integration
AWS Step Functions
Serverless visual workflows that coordinate AWS services.
Official docsOverview
Step Functions lets you compose Lambda, ECS, Glue, SageMaker and 200+ services into reliable state machines using Amazon States Language (ASL).
When to use it
- Long-running workflows requiring retries and branching
- Orchestrating ETL or ML pipelines
- End-to-end test pipelines
Setup
- Console → Step Functions → Create state machine. Pick Standard (long, durable) or Express (high volume, short).
- Author ASL JSON or use Workflow Studio.
- Grant the state machine an IAM role with permission to invoke targets.
How to use
Tiny state machine
{
"StartAt": "Provision",
"States": {
"Provision": { "Type": "Task", "Resource": "arn:aws:lambda:...:provision", "Next": "RunTests" },
"RunTests": { "Type": "Task", "Resource": "arn:aws:lambda:...:runTests", "End": true }
}
}QA use cases
- Orchestrate: provision env → seed DB → run API tests → run UI tests → publish report → teardown.
- Implement retry-with-backoff for flaky tests at the workflow level instead of inside test code.