Adds optimization for related relations

[flovilmart]

So we ran a bit the aggregation optimization in prod and this was not satisfactory, not yielding enough benefits in terms of performance.

Went back to the drawing board and came up with this idea of leveraging the _id managed by mongoDB (which is ascending at creation time) to reduce the amount of keys pulled from the join table.

If you're sorting by createdAt ascending or descending, you'll benefit this optimization, otherwise, that will pretty much be the same.

[montymxb]

Looks to be something up with the test case on postgres, maybe it doesn't know what to do with some of findOptions?

[codecov]

Codecov Report

Merging #4345 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #4345      +/-   ##
==========================================
+ Coverage   92.64%   92.64%   +<.01%     
==========================================
  Files         119      119              
  Lines        8426     8435       +9     
==========================================
+ Hits         7806     7815       +9     
  Misses        620      620

Impacted Files	Coverage Δ
src/Adapters/Storage/Mongo/MongoStorageAdapter.js	93.36% <100%> (+0.02%)	⬆️
src/Controllers/DatabaseController.js	94.84% <100%> (+0.07%)	⬆️
src/RestWrite.js	93.32% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7223add...5bd5cee. Read the comment docs.

[montymxb]

Awesome 👍 ! Also (if you've tested against this) how much of a benefit are you seeing from this change?

[flovilmart]

Instead of pulling data from the full join (possibly 1000’s of objects) this will only pull a few. In only works you’re sorting on createdAt, as it’s represented on the mongoDB_id.

[montymxb]

I got that, I just wanted to know if there was an actual benchmark of sorts that you might have had for a particular case. But that's just for my personal interest.

[montymxb]

Beyond that we're good though!

[flovilmart]

We're seeing very large data in some relations that are actually sorted this way, so that really helps if returning data faster :)