Hello Sonali !
Thank you for posting on Microsoft Learn.
You issue appears to be caused by a transient or backend issue within the service, especially since the same files sometimes succeed upon retry, possibly due to large input sizes, temporary access issues with blob storage, or hitting concurrency limits.
You need to verify if your input files are consistently formatted, accessible, and conform to service requirements.
As a workaround, I advise you to split large batch jobs into smaller chunks and implement retry logic with exponential backoff.
Meanwhile, you can open a support ticket with Microsoft Azure for better assistance.