IMHO this article demonstrates both the flexibility, and a significant issue with PostgreSQL... and that it doesn't offer a good high availability strategy in the box. Progress has been made, and it will likely get there, but it's hard to even suggest using it in many scenarios without an in the box, or at least relatively easy to implement failover solution, with clear, detailed and updated instructions and documentation. And not everyone hase >$100k/year as a baseline for using the listed consulting firms, or can even get a response from half of them.