Check: metrics | ComputeSphere Learn

Goal: Confirm you can read the Metrics tab against a shape's ceiling and tell a busy service from a starved one.

Four questions: three on Module 3, one on Module 1. Every answer explains itself, right or wrong.

A Standard service shows CPU at 0.3 vCPU all afternoon. Should you worry?Yes, it's over the 0.25 ceilingNo, it's well under Standard's 1 vCPUYou can't tell without switching to Max

Two Flex spherelets: CPU on Sum reads 0.3 vCPU, and on Max reads 0.25. What's going on?One spherelet is at its ceiling while the other is nearly idleBoth are over the ceilingThe average spherelet uses 0.25

After a traffic spike, users see errors for a few seconds. The Memory chart dropped to its starting level at the same moment. What do you check first?The deploy log for Image pull failedThe deploy log for Out of memoryWhether CPU was flat at the ceiling

From Module 1: users say pages have felt slow since this morning, but nothing is failing. Which signal answers “how slow, for how many?”ErrorsLatency, as p95 on the Traffic tabTraffic, as the number of requests

Retry as often as you like.