Since your direct question has already been answered, I'll tangent a bit and ask "why do you think you need to back up in the first place?" For those new to Git and _especially_ those who've been conditioned on centralized VCSs, it's not always easy understand the full ramifications (benefits) of a distributed VCS, one of which is that every clone of the repo is a complete clone - every clone is a backup! (although to fully restore "origin" you may need to have a couple of local repos push to it - but you can ALWAYS restore origin from local clones)
If you decide to still run backups, that's fine - but make sure you have a "right reason" for doing it and that it's based on DVCS principles. You're not going to backup to restore to a point in time - VCSs obviously do that themselves; you may not want to backup for disaster recovery purposes, since recovering from local repos may be easier (you almost don't even need to try to recover...whenever the next person does a push from a local, it'll just recreate everything on origin automatically [well...sortof...everything in the history of the commit being pushed]).
NOTE that the same argument doesn't apply to GitBlit config directory!!!