medium10 pts · first answer countsprometheusperformance

Prometheus memory explosion

After a developer added a user_id label to a request duration metric, Prometheus memory usage grew quickly until it crashed. Why?

A.The scrape interval became too short for the number of targetsB.Grafana is caching too many dashboard query results in memoryC.The disk that holds the write-ahead log filled up completelyD.Each unique user_id creates a new series, so series count exploded

← Requests per second in PromQL Cross-account KMS decrypt fails →