Well it's been nearly a full week. I'll be honest, if you haven't found the answer yet to my SLA question, I think it's pretty clear what the answer is. That's not to say that you won't credit your customers if something bad happens- I'm sure AutoDesk will consider things on a case by case basis- that's just good business.
But it does mean that there isn't a public standard that customers can hold you to. Maybe AutoDesk thinks 99% uptime is acceptable? That sounds pretty good right? Well if you do the math, that works out to over 87 hours a year. Hmmm... maybe not so good after all? I do know that the more "nines" you want to support, costs tend to increase exponentially. More likely, you don't actually have an official uptime target for which the system was engineered to meet- that's generally been my experience for things like this. That makes things sorta like the warning you read about the stock maret: "past performance does not necessarily predict future results".
A co-worker just pointed this out to me... an excellent example of why trusting other companies to run your critical infrastructure can have dire consequences:
https://www.theregister.co.uk/2017/02/01/gitlab_data_loss/
<insert "you had one job" meme>
Interestingly, as I write this there is a notice here on the Eagle Forums "Some blogs, ideas, and comments are not available at this time. This should be resolved shortly. Thank you for your patience."