Several issues with logging and concurrency

[dylanmcreynolds]

During a user operations, a test scan was done which took an hour to fail. While it was waiting to fail, prefect server blocked a new_file_832 scan from happening. The following log image shows some interesting points:

1. Though it failed, there isn't that much information (though it did turn out to be a file not found error at the bottom of the log)
2. If transfering to reconstruction files, we should fail the entire flow and parent flows (possibly by raising an exception). There is no point trying to create zarrs or copy anything back.