Inquiry about Handling FAILED Responses in Batch Statistical API

Dear Sentinel-Hub Support Team,

I am experiencing an issue similar to the one described in the following URL:

Given that this issue may occur, I need to verify all the JSON files downloaded to a specified S3 bucket by the Batch Statistical API to check for any responses marked as FAILED. If any are found, I would need to create a new request for those specific responses. This process is time-consuming due to the need to read files from S3, so I would like to avoid it if possible. Is there any way to know in advance if a response has FAILED?

Additionally, if a significant number of polygons have FAILED, it could result in a substantial consumption of Processing Units (PU). In such cases, is it possible to get a refund for the consumed PU?

Thank you for your assistance.

Best regards,