Why is everyone skeptical about MongoDB?

Question

MongoDB

Good afternoon.
I meet a lot of questions like "Which is better to choose for my site / web service NoSQL or RDBMS?". Constantly advises either MySql or PostgreSQL. Everyone is afraid to use MongoDB as their main database. I'm not saying that Mongo is a pill for all ailments, but there are many tasks that can be given to it, but instead everyone runs from it to the above databases.
The question is as follows. Why is everyone skeptical about MongoDB?
UPD. Edited the question.

Answer 1

Retrogrades are skeptical, people who are afraid of everything new and unfamiliar.
NoSQL has its own problems and advantages, you need to understand when and how to use it. Many people try to use them the old fashioned way, use their crutches to organize the relational model, and then are surprised that their solution does not work. Mongs are made for something else. They are good for storing a lot of relatively small data structures and retrieving them quickly. And all this should scale well. For example, a telephone directory for a trillion numbers, which by no means fits on one machine, while someone constantly writes new numbers into it and there must be replicas to it in several data centers. Plus you should have full text search across multiple fields in multiple languages.
It is possible to do this on MySQL, but the solution looks more complicated, but it also has its advantages. For example, you can rename all Vasily's with one simple movement in the form of editing one entry. In Monga, you will need to change all entries with Vasily.
Both solutions have replication, sharding, etc. Mongu is a little easier to set up, that's all.
Monga is not suitable for those who need a clear commit of transactions, because there data integrity is based on the flow of events.
You just need to carefully read on what bases to use for what.

Answer 2

your title is different from the question
1) NoSQL vs RDBMS - one
2) MongoDB for * we are Memcached, Redis, Aerospike, other databases - another
answer is about here: habrahabr.ru/company/jelastic/blog/166845
there is an article, year 2012, either from Percona or from MariaDB, where it is shown how to adapt them to NoSQL and achieve the same speeds, the alignment is simple: you turn off SQL features - you get speed))
for the convenience of SQL, they also do not stand still, Json was recently added in Muskul, accelerated ( a 30% performance increase is claimed in version 5.7), etc.
well, then it’s simple: if they are the same in terms of convenience, then how will they lead with tens of millions of records and a comparable number of transactions? Replication, again? Muscle and Postgres have a long background in these matters, but Mongo?
Therefore, Mongo is perceived as a fast storage for non-critical data, and is used in the wake of the classic RDBMS

Answer 3

In my opinion, each database can find its application. It all depends on the need, the capabilities of the database systems, as well as the basic requirements of the project. The first two from your "list" are familiar databases, right ... they are often used by many, they are simple in logic and easy to "use". There are also N-th number of things that cannot be done in MariaDB (Mongo DB), in my opinion, difficulties arose with sorting and more complex select queries. If my memory serves me, then Mongo DB is faster.
Why not use? Yes, most likely you are right - this is a matter of habit for most "homemade" ones. As for fear, I'm not sure, many simply may not know what kind of animal this "Mongo" is...