Hacker News Reader

Hacker News

Comment by b0afc375b5 | Hacker News Reader

Parent Full thread

b0afc375b5

For anyone else curious, I found the postmortem here:

https://mcfunley.com/why-mongodb-never-worked-out-at-etsy

Which was compiled here:

https://github.com/icy/w2w

View on HN

That repo is interesting. A quick ctrl+F seems to indicate that pretty much every instance of "MongoDB" is "Moving from Mongo to Postrges or DynamoDB" (there is one single entry of moving to Mongo from MySQL). Almost as if Mongo is just not a good database (or people are too eager to use it for things which it does not do well).

Making efficient use if mongodb is very difficult but if you build your app and expectations correctly you can get something very performant. For example pre-4.x listing huge collections was unexpectedly extremely slow.

Yeah Mongo is 10 years old or so at this point. This article was written in 2015 about decisions made years earlier. It's now reached maturity and stability. It's now "boring tech".

It may be more established but it suffers from a similar problem: some people continue choosing it for the wrong reasons, such as "our backend is in JS, most of our devs only know JS, and it's easy to just dump objects in there", which is an abhorrent reason for choosing it, and will end up biting you.

IMO this would be a case where if you're dealing with a relational domain and the engineers really don't know SQL you should either (a) rethink your hiring policy or (b) spend one of your innovation tokens in having everyone learn SQL.

(I have to add the inevitable disclaimer that I actually love JS and do not want my words to be misinterpreted as a cheap dig at it)

Does mongoDB still use math.random to select 10% of errors and only report them, while 90% are ignored? Or did they correct this "behavior"?

Huh? That sounds like crazy talk..

Um, even if it's mature technology that doesn't negate the costs of operating two different databases which is a large part of what the original presentation is about.

Reply on news.ycombinator.com