serious reliability problems that you have to engineer around yourself Sounds like AWS - they explicitly state that you can't count on their stuff and you need

serious reliability problems that you have to engineer around yourself Sounds like AWS - they explicitly state that you can't count on their stuff and you need to engineer your stack that way.