As a big advocate for NodeJS, the association of the stack to Mongo kills me. Mongo legitimately has no place in this world. If you truly have a case for schema

As a big advocate for NodeJS, the association of the stack to Mongo kills me. Mongo legitimately has no place in this world. If you truly have a case for schemaless, you have a million better options, and from most use cases I've seen, it's Postgres's JSONB. You get schemaless JSON, same as Mongo, with a GIN index to index every field in the column. But then you also get transactions, and actual columns with actual constraints for the cases where you may undoubtedly muster the commitment. You also get SQL... And if you're talking data sets that aren't going to work with an RDBMS, it's still the wrong tool for the job. You're going to end up in Kafka/HDFS/S3/Cassandra/etc, with some custom indexing/caching strategy in MapReduce/Redis/ElasticSearch/etc. It is literally impossible for me to grant Mongo the slightest consideration as the appropriate tool in any single part of a reasonable stack.