Currently the files added in the archive are compressed which is resource intensive for the server. We should disable the compression to reduce the load (FWIW this was part of the acceptance criteria of
story #12404).
There is also an opportunity to decrease the I/O load by not reading multiples times the files that are added to the archive.