I have used MongoDB and MySQL, and for the last two years, I am using CockroachDB, which is better than these two. I have recommended it to two more of my colleagues who have also started working on CockroachDB, and they are building an application for the global market. They have started using CockroachDB because it has been very helpful.
I am using Hasura just for the GQL, GraphQL.
Hasura has connectivity with CockroachDB, and I am not directly calling CockroachDB, I am just calling Hasura. Internally, Hasura calls CockroachDB, gets the data, and populates it into our application. It serves as middleware.
We have real-time subscription in Hasura.
If I change anything in CockroachDB, it immediately impacts Hasura. For example, if I add new fields, it gets reflected in Hasura.
Integrating microservice architecture with Hasura provides an endpoint to incorporate into your microservices. It works as a connection string, and that's its main role.