intermediate · latency
The Endpoint That Got 20× Slower
An endpoint went from 100 ms to 2 seconds. How do you find out why?
Practical scenario
GET /orders/{id} has a p95 of 2.1 s since roughly 09:00. It was 100 ms yesterday. Traffic is normal, error rate is normal, and three people are already suggesting fixes in the incident channel.
What it tests
- • Whether the candidate measures before theorising
- • Whether "what changed?" is their first question
- • Whether they can localise time to a layer rather than guessing at fixes