Intro to PostgreSQL Caching layers: App vs. Proxy vs. Database

I haven't seen too many comprehensive approaches to this topic, so I took a swing at it.

Goal isn't to be exhaustive, but to be thorough enough to help people trying to get their bearings on the topic.

It covers the core differences and trade-offs of different cache flavors:

1. In your application layer
2. At the proxy level
3. In the database

Did I miss anything?

Edit: I'm curious what everyone's actual production stack looks like for this. I dropped my mental model on the cost vs. complexity trade-offs in the comments below. Wondering if people actually handle read optimization that way.