問題文
The same team is deciding what to measure for the order API. They want the smallest set of measurements that still tells them whether the service is healthy for its users. Which set should they start with?
選択肢
- The average processor utilization of each host, because saturation is the root cause behind every user-visible problem in a request-serving system.
- The number of log lines written per minute, because the volume of logging on the replicas tracks how much work the service is actually doing.
- A single overall health value between 0 and 1 computed inside the application, so that dashboards and alerts only ever need to read one series.
- The number of requests served, how many of them failed, and how long they took, each broken down by a small number of labels.