Simplify reader usage in consumeGzipUpload
The following discussion from !2134 (merged) should be addressed:
io.Copyexists because it was likely copied and pasted from
consumeZipUploadwhere thw temporary copy is required because opening a
zipfile requires knowing the file size.
For Gzip, I think we can just replace this with
gz, err := gzip.NewReader(reader), reading directly from the io.Reader passed to the function without the temporary buffer. `