When the SDK throws the generic Request could not be completed error on specific PDFs, suspect the default timeouts before the documents. Build a ClientConfig with explicit timeouts - e.g. 10s connect and 40s read - and pass it into the ExecutionContext alongside your credentials. The reporter's files parsed fine once the read timeout was raised; the SDK just gave up too early and wrapped it in a meaningless exception.

Context: Stack Overflow 68372461 (accepted answer, 2 votes): Adobe PDF Services threw a generic SdkException - Request could not be completed - on some company report PDFs with no useful detail. The accepted answer found the files were fine in Node and diagnosed it as a timeout, not a document problem.