When configuring an EMR Serverless application, you can choose the disk size for your preinitialized capacity and a maximum disk limit for the application:
However, at least for Hive jobs, EMR Serverless stores both the scratch and warehouse directories on S3. So what is this configured disk space used for? I've searched the documentation and there's no mention of how this space is used, or why it's needed.
This does answer the question, thanks. But we needed this answer 3 months ago. We've since moved our company's infrastructure to a new cloud provider.