Every now and then an upload will fail processing. Fair enough, software is not perfect and errors can occur.
In many cases this is not through the fault of the file, as performing a re-upload will then work successfully. I will admit that sometimes a re-upload is the only resolution in the case that the initial upload may have been corrupted.
But, there are other instances such as sharing/consuming packages and clash checking, where this is entirely on-platform using known good/working models. This shouldn't require a re-upload, or repeating the same cross-module workflow task to get a successful result.
I realise that performing the same task and expecting a different result is madness. However, in all cases, I would like the ability to at least re-try the process at least once. If that re-try also fails, then it's fair enough that additional work is required to repeat the entire workflow.
As a side note, it would also be helpful if the processing errors came with some more details, especially if it's an underlying issue with the file that requires more significant work.