AFAIK they have sharded their pages and repositories in a way that if one gets DDoS, usually it's only that piece that gets downtime (I know this because there was a problem with the Chinese government and a Github repo, that got constantly attacked, but that was the only repo suffering).
Ofcourse, if one does a sharding, it's bound not to be repo by repo, so there might be adjacent ones suffering.
But I use GH everyday, can't remember a time when it was offline.
Here is the confirmation of such a mechanism (they simply rate limit certain repos):
"All of these factors combine to make CocoaPods/Specs one of the top five most resource-costly repositories that we host on all of GitHub.com. And that is why it is rate-limited; otherwise it would consume even more resources and cause service interruptions for other GitHub users. The symptoms of the rate limiting for you and your users are that your repository accesses (clones, fetches, pushes) have to wait in a queue on our end, sometimes for a long time, before being processed. This causes fetches/clones to take much longer than they would otherwise, and might cause timeouts at your end. Moreover, if the load on our servers becomes too overwhelming, a fraction of the accesses might be rejected altogether."
That uptime measure is for the application server, which I suspect does not include GitHub Pages which is hosted separately on static content servers.
Lastly, 99.9% uptime over a month means 43 minutes of downtime. To me, that sounds like quite a lot.