Two things which might help:
-
The more machines there are which back up to a specific machine, the more important it is going to be to have the option to take, for example, hourly snapshots not on the hour, but rather to take them an hour after some random time since the machine in question started (and then to continue every 60 minutes after that). That’s the easiest way to spread the load on the machine hosting the repositories and the network infrastructure closest to it.
-
Machines which move across networks will need a way to disable kopia-UI’s knowledge of repositories which cannot be reached at the moment. See my two posts:
- Different approach to handling multiple repositories which explains this in greater detail and gives a few ideas about solving it longer term.
- My reply, which shares a script I wrote to deal with the core problem and which is working perfectly for me.