Use Case

As a... Node.js driver user
I expect... the performance to remain roughly the same between releases
So that... I get a consistent responsive experience

Reproduction

https://mongodb.slack.com/archives/C03MPA2FY6S/p1721991547262229

import { MongoClient } from 'mongodb';

const client = await MongoClient.connect('mongodb://localhost/');
const cursor = client.db('test').aggregate([ { $documents: [{}] }, { $set: { field: { $reduce: {input: [...Array(20).keys()], initialValue: [0], in: { $concatArrays: ['$$value', '$$value'] } } } } }, { $unwind: '$field' }, { $limit: 1_000_000 } ]);
await cursor.toArray()
await client.close();

(run with time node test.js).

User Impact

• Cursor iteration is unexpectedly slower

Dependencies

• Cursors, On Demand BSON, Driver Benchmarks

Unknowns

• What is the main bottleneck introduced?
• What is the remedy to the regression?

Acceptance Criteria

Investigation Requirements

• Investigate performance before and after the identified regression commit
  • Collect flamegraphs and cpuprofiles for comparing
  • Present a theory and suggested patch to the major bottleneck(s)

Implementation Requirements

• TBD, See above.

Testing Requirements

• Investigate the performance alerting system. Did our benchmarks identify this and not warn us? Are we missing a certain kind of benchmark to identify this?

Documentation Requirements

• None

Follow Up Requirements

• None