On-callMediumoc-g257

Subject P99 regressionLevel Mid–Senior~30 minCommon in Reliability & on-call interviewsIndustries Technology, Software development

Question

Your read path is normally a 2ms cache hit, but a few times an hour p99 spikes to 1.5s and the origin database briefly sees a 50x read burst, then everything returns to normal. The bursts line up exactly with the TTL boundary of one very hot cache key (a global config/leaderboard blob with a fixed 5-minute TTL). The cache and DB are otherwise healthy and there was no traffic change. How do you triage and fix?

What a strong answer looks like

Stop the bleeding first (mitigate), then form hypotheses from real signals. Separate root cause from symptom, communicate status as you go, and close with what prevents a repeat.

Learn the concepts

Diagram & narrate the incident

Loading whiteboard…

Run or narrate your approach, then ask the coach.