Your endpoint returns 100 orders.
It then fetches the customer for every order.
Congratulations: one innocent endpoint just created 101 database queries.
The N+1 problem hides well in ORMs because the code looks clean while the database performs repetitive work.
Use query logs and traces. Load relationships deliberately with joins, relation loaders or batched queries.
Performance starts with knowing how many queries one request actually executes.
What is the highest query count you have found behind one endpoint?
#NestJS #TypeORM #DatabasePerformance

Smiley Face

Welcome to the NestJS Developers Hub, a collaborative community for developers passionate about NestJS! Whether you're a seasoned expert or just getting started, this group is dedicated to sharing knowledge, solving problems, and advancing our skills

Nestjs Developers

The N+1 query problem occurs when an endpoint fetches a list of records and then issues a separate database query for each related record — turning one request into dozens or hundreds of queries. ORMs mask this because the code looks clean. The fix is to use query logs and traces to measure actual query counts, then load relationships deliberately using joins, relation loaders, or batched queries.

Day 9/30 — Kill the N+1 Query

The code needs to be checked. Most of the iterative work to fetch data can be done with nested joins. Also, you can use the
<pre><code>SELECT x FROM y WHERE y.id IN(id1,id2,...)
</code></pre>

Query count is a great regression target, not just a debugging observation. I would add an integration test that loads 1, 10, and 100 orders and asserts the number of database round trips stays bounded rather than growing linearly. Joins are not automatically free, though: one-to-many eager loading can multiply rows and break pagination. A batched loader plus per-request query tracing often gives a safer balance between round trips, memory, and result shape.

Put query-count traces at the serializer boundary, where an efficient repository call can quietly become one round trip per nested field

Please share an example of N+1 query problem and provide solution to address it.