Why should you care?

At Diwali sale scale, a single missing index or unbounded $lookup can take down checkout. Tuning is a checklist, not a one-off createIndex.

See it live — copy this example

Open mongosh or MongoDB Compass, select database nosqlverse, then run the example. Change one field and run again.

// Practical tuning checklist in mongosh
db.serverStatus().connections
db.currentOp({ "active": true, "secs_running": { $gt: 2 } })
db.orders.aggregate([
  { $match: { tenantId: "acme", placedAt: { $gte: ISODate("2026-07-01") } } },
  { $group: { _id: "$status", n: { $sum: 1 } } }
], { allowDiskUse: true })
db.orders.createIndex({ tenantId: 1, placedAt: -1, status: 1 })

Run Example »

Edit the code below and click Run to see the result in Toolliyo’s live editor.

What happened?

• connections shows pool pressure.
• currentOp finds slow operations.
• The aggregation is bounded by tenant and date.
• allowDiskUse helps heavy groups.

Practice next

1. Capture top slow queries from Atlas Profiler or currentOp.
2. Ensure working set fits RAM for hot collections.
3. Cap aggregations with $match early and allowDiskUse intentionally.
4. Run the same report with readPreference secondary (replica lesson).
5. Add maxTimeMS to risky queries.

Remember

Tune queries and indexes before scaling hardware. Watch currentOp and connections. Isolate heavy analytics from OLTP primaries when needed.

National e-commerce peak

Secondary nodes serve analytics; primary keeps checkout writes; indexes match sale queries.

Outcome: Store stays up while finance still gets reports.