Thanks Joao. That is a helpful document. I've read that one and most of the other related doc in the Evaluation guide and developer documentation.
As that document points out, an admin (human) has to manually select additional capacity in order to achieve scaling up or out. I'm not under-valuing the work that the native OS platform does to 'just take care' of the rest of the devops / infrastructure details, but what if I wanted my application to be able to automatically scale based on monitoring triggers and/or programatically? That is, dynamically without human intervention.
I like the native management the OS platform provides so I don't want to have to use containers unless it's absolutely necessary.
See?