MongoDB Performance Optimization Techniques

Introduction

MongoDB is a powerful NoSQL database, but without proper optimization, queries can become slow as your data grows. In this article, I'll share techniques I've learned while working on enterprise-scale applications at Platform Science.

1. Indexing Strategies

Indexes are the foundation of query performance in MongoDB. Without proper indexes, MongoDB performs collection scans, which become increasingly expensive as data grows.

Single Field Indexes

Start with indexes on fields you frequently query. For example, if you often search by userId, create an index:

db.activities.createIndex({ userId: 1 })

Compound Indexes

For queries with multiple conditions, compound indexes are more efficient than multiple single-field indexes:

db.activities.createIndex({ userId: 1, timestamp: -1, status: 1 })

2. Query Optimization

Use explain() to analyze query performance and identify bottlenecks:

db.activities.find({ userId: "123" }).explain("executionStats")

Key metrics to monitor include totalDocsExamined vs nReturned. Ideally, these should be close to equal.

3. Aggregation Pipeline Best Practices

Order your pipeline stages efficiently. Place $match and $limit early to reduce the document count before expensive operations:

db.activities.aggregate([
  { $match: { status: "active" } },
  { $limit: 1000 },
  { $lookup: { ... } },
  { $group: { ... } }
])

4. Bulk Operations

For inserting or updating large datasets, use bulk operations instead of individual writes. This reduces network round trips significantly:

const bulk = db.activities.initializeUnorderedBulkOp();
documents.forEach(doc => bulk.insert(doc));
await bulk.execute();