Java/Spring/JVM · Production incident
Every endpoint is slow at once
Every endpoint of a service is slow at the same time while its CPU is idle. How do you contain it?
Thursday 14:12 · storefront-apiIllustrative scenario, not a real incident
storefront-api: p99 above 9 s on every endpoint, 38% of requests end in 503 at the gateway
storefront-api is a Spring Boot service on Java 21. It serves product pages, prices and the basket for the web shop, at about 300 requests a second.
Since 14:05 every endpoint has been slow, the price and basket endpoints included. CPU on the pods is at 14% and memory is flat. storefront-api itself was last deployed on Monday.
What the dashboards and logs show
- p99, all endpoints
- 9.4 s (120 ms an hour ago)
- 503 at the gateway
- 38% of requests
- CPU on the pods
- 14%, lower than an hour ago
- Deploys of storefront-api
- none since Monday
Clarify · optional · scored under "Hypothesis before change"
Ask the interviewer
0 of 3 asked
Each question reveals a fact. What you ask before you act is part of your score.
First move · scored under "Mitigate before root cause"
Shoppers cannot see prices or their basket while you investigate. What do you do first?
Your rubric score, the model answer, the follow-up questions and the related lessons appear once you have answered the last step.