I’m using restic too for part of my backups, having moved from duplicity. My experience has been (on a dataset which is relatively compressible ie not audio/vid

I’m using restic too for part of my backups, having moved from duplicity. My experience has been (on a dataset which is relatively compressible ie not audio/video files) that deduplication has saved me quite a lot more space vs duplicity than I lost through not having compression.