I've always hated MongoBD, but here's the thing.<p>I think I have the perfect use case for a MongoDB cluster, perhaps k8. What if you have a super-normalized relational database, the source of truth, and your life depends on keeping it alive so you make backups every ten minutes, but what if it's too expensive to query from the front-end? What if you wanted to remedy that? What if, each night, you flattened out your relational DB into ten-ish document collections that are super swift to read from?<p>>> But what about the customers?<p>What about them? Sure, there's a theoretical chance something got lost on its way from being a virtual document composed of data from fifteen different database tables into becoming a single materialized MongoDB document, but the loss is small compared to the savings you make from not having to scale up your SQL.<p>>> But what about Elasticsearch?<p>But here's the thing, I've always hated Elasticsearch.