This App Broke with Just 10 Users — Here’s How I Fixed It

In this video, I break down how I fixed a severely slow Django + GraphQL + Next.js application that became almost unusable with just 10 users.

At its worst, a single request could take up to 10 minutes.

After digging into the root cause and using AI the right way, I was able to reduce:

* 15 requests down to 2
* 1,345 SQL queries down to 8
* 5,549ms down to 235ms

The biggest issue was the GraphQL N+1 problem, along with unnecessary frontend requests and missing observability in the system.

In this video, I show:

* How I identified the real bottlenecks
* Why telling AI to “make it faster” is not enough
* How to build observability first
* How I used AI to optimize both backend and frontend
* What to review with your own engineering judgment before applying changes

This is the exact process I used to dramatically reduce latency in a real-world application, and you can use the same approach in your own projects.

#django #graphql #nextjs #webperformance #softwareengineering #ai #programming #backend #frontend #optimization

Видео This App Broke with Just 10 Users — Here’s How I Fixed It канала Harith Bakhrani