I disagree. I think there are some good use cases for MongoDB.
For example, applications like Strapi where the user can create their own data structures at runtime are a great fit for MongoDB.
Also for Logging, while you could use RDBMS, it will only work if your data structure is fixed. for SaaS applications like ServerDensity responsible of log aggregation of many clients which can have logs in multiple formats, a Schemaless Datasotre like Mongo is perfect.
Activity feeds, Analytics, IoT data and all the cases where the Data Schema is not known in advance or changes very rapidly.
Some important things to remember about RDBMS. Writes and Joins are in general expensive.
I agree with you on the MEAN stack. I think "Mongo and Non relational by default" is probably not the best fit for common use cases.
But one fact that I believe, makes more people go to that route, is SaaS like Firebase, Dynamo DB etc which have a great free tier, thus making then perfect for smaller applications, personal projects or MVPs.
I dont know any "RDBMS as a Service" that is free to use. Services like Amazon RDS or Google Cloud SQL are very expensive.
The only one that is close is Heroku Posgres but its tied to Heroku.
In conclusion, while I believe Non Relational is being abused a little and its not the best fit for many common use cases, Databases like MongoDB definitely have its place and are quite useful in some contexts.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.