I'm not sure I understand your point?
Even if I ran it on my own hardware, where I was responsible for plugging the cable in, it would still have the same single point of failure problem in the event that my ISP started suffering Network issues. I've never had an ISP that didn't have some sort of issues at some point.
Alternatively, if you think I am blaming my cloud provider for being the single point of failure, please read again. I specifically mentioned that I want to fix up the application so that it is highly available. I entirely and completely blame myself for the current situation. I didn't build it (otherwise it would be 12 factor[1] from the start and horizontally scalable), but I maintain it now.
[1]: https://www.youtube.com/watch?v=REbM4BDeua0