hi nick,
That's what is supposed to happen- your job did fail. I would suggest creating different jobs for each client (or small groups of clients if you have to). That way your failures will more accurately reflect the state of your environment.
Of course, you still need to fix the broken client.
And You can configure policy based backups to better manage your backups, this will also help you to pinpoint the exact host being failed easily.
Hope this helps
---Cheers
shirmal