Every time someone makes the claim that downtime should be a warning sign about going with a PaaS provider (or, indeed, an IaaS provider, or in some cases, people even make this claim about going with someone else's Data Center) - I always respond: "And why do you believe that you would do any better?"

Every environment I've been involved in as an operations professional for the last 15 years has experienced downtime regardless of how much we invested in staging, testing, change control procedures, redundancy and ITIL methodology.

While it is true that there are some environments in the world that experience no, or close to no downtime (The NYSE, google.com, the Space Shuttle) - it's usually not worth the investments those organizations make in ensuring 100% uptime.

I'm more interested in seeing what the monitoring methodology is, what the response protocols are, and what the reported downtime has been over the last year has been, than I am about getting too concerned about the occasional outage.