While I'd agree with the general premise that diversification is a good thing in platform use if high-availability is a requirement, given that this outage was single-AZ, this particular outage should really highlight the point that your application should be multi-AZ scaled if it needs to be up.