I am struggling with how often I should perform backups. How frequently do you all back up your data?
Currently, I have an automated process where my VMs are backed up daily while I sleep, and then I manually copy those backups to external storage when I wake up. However, because the file sizes are so large, this process has started to become a hassle.
I would like to hear your opinions and use them as a reference. Please let me know what you think.
Personnally, i use wyng for all Vms from @tasket .
It’s very easy and it’s incremental so it’s create backup very quickly 
edit: for dom0, i use qubes-backup because wyng doesn’t backing up it.
I dont use Wyng or Qubes-backup, except in specific circumstances.
For data backups, I run backups according to the type of qube, and how
often the data changes. After first full backup, all are incremental.
For important qubes, I backup daily.
For very important qubes, I back up hourly.
I think there might be some answers here:
Do you use the default compression?
I only use the default Qubes backup monthly or just before upgrading to a new release.
I have made my minimal Qubes OS setup based on a script so I can easily recreate all my qubes. Therefore, I only backup important files and folders from my home directories. The backup itself is fully automated and takes 1-5 mins.
https://codeberg.org/ERROR404/qubes-os-dm/src/branch/main/06-backup.sh
Feel free to reuse, copy or modify.
I have a bash script that I run to shut the machine down each night. It starts sys-backup which mounts my raid fileserver and then checks each VM timestamp to see if it had been run that day, and if so it runs the qubes command line backup utility for that VM.
Each VM is backed up into its own individual archive in a dedicated directory on the file server for that vm. This way I can manage exactly how many copies of each vm’s backups I want to maintain, and I don’t need to search each archive to see which has the latest version of what VM. Once the backup is done an integrity check is done, and when all is complete and verified it locally mails me the final report and then shuts the machine down. Alternatively I can have it run and back up only VMs that are not currently running and allow me to just continue working.
If any VM ever needs to be recovered I just go to that dedicated directory on the file server and restore the one archive having the latest timestamp, or any other version as far back as I choose.
Prior to backing up dom0 I have it run a script to create a quick tar archive in the users home directory that contains all the specific machine configuration. Expanding that archive puts everything back in place on the new machine. When restoring dom0 after a major upgrade all my scripts and configuration will be immediately accessible, and those utilities help to restore all the rest of the machine.
Why not just have the backups created on external storage directly? Or, if you want copies in both places, automate the copy part too.