Am I the only one who thinks these "repeatable workflows" are just a fancy way to lock you into an expensive toolchain? Everyone posts these elaborate pipelines with 5 different SaaS tools.
My issue:
* The second you scale, the costs explode. No one talks about that.
* They're never optimized. You're paying for idle time, over-provisioned compute, and redundant storage.
* Zero pricing callouts. It's always "just use this managed service."
Example: A workflow that runs a Lambda to trigger Step Functions, processes files in S3, uses Comprehend, then writes to DynamoDB. Looks neat on a diagram. Now run it 10 million times a month. The bill will make you weep.
Where's the workflow for cost-aware design? Where are the recipes for:
* Spot instances for batch jobs
* Reserved Capacity for core services
* Turning things OFF when not needed
show me the bill