Introduction

MongoDB is powerful but easy to misuse. Here's what actually matters in production.

1. Always Use Indexes

Without indexes, every query does a full collection scan.

// Create index on frequently queried fields
db.users.createIndex({ email: 1 }, { unique: true })
db.posts.createIndex({ userId: 1, createdAt: -1 })

2. Use Projection to Limit Fields

// ❌ Bad — fetches everything
const user = await User.findById(id)

// ✅ Good — only what you need
const user = await User.findById(id).select('name email avatar')

3. Paginate Everything

Never return unlimited documents:

const posts = await Post
  .find({ published: true })
  .sort({ createdAt: -1 })
  .limit(10)
  .skip((page - 1) * 10)

4. Use Aggregation Pipeline

For complex queries, aggregation is much faster than multiple queries.

5. Connection Pooling

Always use a singleton connection pattern — never create new connections per request.

6. Validate Data with Mongoose

const UserSchema = new Schema({
  email: { type: String, required: true, unique: true, lowercase: true },
  age:   { type: Number, min: 0, max: 120 },
})

Conclusion

These practices will make your MongoDB database reliable, fast, and maintainable at scale.