> in 2016 that doesn't take much: spin up two instances in different AWS datacenters and fail between them and you have Disaster Recovery
Things that look simple on the surface are often not easy to implement in practice - especially when you're not starting with a green field.