🤓 Does your Lambda function call Service A, check the response, retry with delays if it fails, then call Service B, while writing the current state to DynamoDB after each step?
Great point! Durable Functions work well for simple sequential flows with callbacks - no 15-min timeout or charges during waits.
You're absolutely right that the line is questionable. My take: if you're writing orchestration logic in code (state management, error paths, retries), you're likely past that line. For complex workflows with branching, parallel execution, or multiple approvals, Step Functions saves hours on visual debugging alone.
By the way it's now possible to simulate Step Functions like workflow with a new feature called "durable functions".
https://aws.amazon.com/blogs/aws/build-multi-step-applications-and-ai-workflows-with-aws-lambda-durable-functions/
You can invoke another Lambda & wait for a callback w/o being charged/limited by the timeout of 15m.
While it's still questionable on where is the line between creating proper orchestration with SF and doing the low-effort flow with Lambda.
Great point! Durable Functions work well for simple sequential flows with callbacks - no 15-min timeout or charges during waits.
You're absolutely right that the line is questionable. My take: if you're writing orchestration logic in code (state management, error paths, retries), you're likely past that line. For complex workflows with branching, parallel execution, or multiple approvals, Step Functions saves hours on visual debugging alone.