I don't know, I think heroku should provide the value proposition that - hey we have it covered, if we rely on another cloud it's on use to build in redundancy

I don't know, I think heroku should provide the value proposition that - hey we have it covered, if we rely on another cloud it's on use to build in redundancy and reliability atop it so as not to burden you, the app developer who is paying us to take care of operations and scaling. What would be interesting to me about an Amazon outage being behind a heroku outage would be to keep a tally, and if heroku didn't manage to build in more reliability to be resilient to even an amazon outage in a particular region, to question whether they were a good fit.