Backend · Mid-level · Theory

After a release, the API's 99th percentile response time went from 200 milliseconds to two seconds, while the average barely changed. How would you track down the cause?

~4 min by voice

How to approach it

This checks whether you understand that p99 growing while the average stays flat is a problem in the tail of the distribution, and whether you can localize it systematically: first limit the damage, then narrow the area with metric breakdowns, tracing and logs, and check the typical resource causes of tail latency.

Have you seen this question in a real interview?

Answers0

Avg. score—

Strong—

Practiced0

Topics

observability tracing debugging

Similar questions

A FastAPI service started responding with delays of several seconds under load, even though the CPU is barely used. The endpoints are declared with async def. How would you find the cause? Mid-level What's the difference between logs, metrics and tracing, and what is each of them for? Junior You have microservices, and users complain that requests sometimes take ten seconds, even though average latency looks normal. How would you look for the cause? Senior A user reports a bug in production, but you can't reproduce it. What do you do? Mid-level In production, a service crashes with an OutOfMemoryError every couple of days. How would you go about finding the cause? Mid-level

Answer by voice

The interviewer will ask this question and listen. Think out loud and ask for a hint if you need one. When you're done, pause or tap “That's my answer”. In about a minute, your score and the points you covered show up here.

Best with headphones · ~4 min