I had some positive outcomes using Postgres and Mongo as essentially a front end cache. The use case was unique in that all of the spark/etl jobs ran against po

I had some positive outcomes using Postgres and Mongo as essentially a front end cache. The use case was unique in that all of the spark/etl jobs ran against postgres, which then triggered caching workers to build up a more document oriented cache that the front end talked to. Essentially allowing the front end to get a KV on steroids and documents already prepared well to be rendered and pre joined. Against a specific use case. I was never burned by Mongo, but I wouldn’t choose it if I had to have one backend only. It lends itself nicely as part of an ecosystem imho