Default upload behavior should not cancel already consumed blobs

With the fix done for NXP-24979, each time you upload a blob with a nuxeo-dropzone it tries to clean up the transient storage even if the blobs are already consumed (i.e. moved to their final destination) which produces a couple of errors and 404 in the browser console.

If the bash is already complete but never consumed, then the transient storage will eventually be cleaned up by server workers anw.