medium10 pts · first answer countskubernetescost-optimization

Cutting Kubernetes cost safely

A Deployment requests 4 CPU per pod, but its p95 CPU usage is 0.6 CPU and peaks at 1.2 CPU. The cluster bill is high. What is the best first step?

A.Remove all CPU requests so pods are BestEffortB.Keep the requests and add more nodesC.Set the request to 0.6 CPU and a hard limit of 0.6 CPUD.Lower the request to about the observed peak plus some headroom, then watch latency and throttling

← Catching certificate expiry Take the medium mock interview →