How to Optimize MongoDB Queries With Indexes

How to Optimize MongoDB Queries With Indexes

Slow queries kill UX. Here is how to optimize MongoDB queries with indexes.

Step 1: Find Slow Queries

Use the MongoDB profiler or your APM to find slow queries. Look for queries that scan many documents (high totalDocsExamined).

Step 2: Run explain()

User.find({ email }).explain('executionStats');

Look at:

• executionStats.totalDocsExamined: should be small (just the matching docs).
• executionStats.totalKeysExamined: should be small (just the index entries scanned).
• winningPlan.stage: COLLSCAN means collection scan (bad); IXSCAN means index scan (good).

Step 3: Add an Index for the Query

userSchema.index({ email: 1 }, { unique: true });

Re-run explain(). totalDocsExamined should drop.

Step 4: Use Compound Indexes for Multi-Field Queries

connectionRequestSchema.index({ fromUserId: 1, toUserId: 1 }, { unique: true });
messageSchema.index({ connectionId: 1, createdAt: -1 });

Order by the ESR rule: equality, sort, range.

Step 5: Use .lean() for Read-Only Queries

const users = await User.find().lean();

Skips hydrating full Mongoose documents. Faster and uses less memory.

Step 6: Select Only the Fields You Need

const users = await User.find().select('firstName email').lean();

Reduces payload and memory. If the index covers the fields, it is a covered query (even faster).

Step 7: Paginate With Cursor Instead of Skip

Skip is slow for large offsets (MongoDB still scans skipped docs). Use cursor-based pagination:

const messages = await Message.find({ connectionId, createdAt: { $lt: cursor } })
  .sort({ createdAt: -1 })
  .limit(20);

Step 8: Avoid $where and $regex on Unindexed Fields

$where runs JavaScript (slow). $regex on an unindexed field scans the whole collection. Use a text index for search, or index the field if you need regex.

Step 9: Use countDocuments Carefully

countDocuments can be slow without an index. If you only need to know if documents exist, use findOne and check for null:

const exists = await User.findOne({ email }).select('_id');

Step 10: Monitor in Production

Watch slow queries with the profiler or APM. Add indexes as query patterns emerge. Do not assume your indexes are perfect; data and queries evolve.

The Takeaway

Optimize MongoDB queries: find slow queries with the profiler, run explain(), add indexes, use compound indexes for multi-field queries, use .lean() for read-only, select only needed fields, paginate with cursor, avoid $where and unindexed $regex, use findOne for existence checks, and monitor in production.