← ☸️ Containers & Kubernetes
Advanced
CFS quotas and throttling: what a CPU limit really enforces
A CPU limit bounds CPU bandwidth over an accounting period, so runnable work is throttled the moment the group or an ancestor exhausts its quota. That produces latency spikes while average CPU looks comfortable, and the page gives the arithmetic, the diagnostic reading of cpu.stat, and the conditions under which each of the three remedies is the right one.
View Premium accesscourse lessons, concepts and answers · see current terms
RELATED CONCEPTS
PRACTICE THIS IN REAL QUESTIONS
Incident Response & Production Debuggingp99 spikes, probes time out, CPU utilisation reads 30 percent and nothing is OOMKilled. cpu.stat shows nr_throttled climbing. Explain it.→Incident Response & Production DebuggingA JVM pod is OOMKilled with the heap at 60 percent of its maximum. The container limit is 4Gi and Xmx is 3Gi. Explain it.→Containers & KubernetesDraw me a Kubernetes cluster. What are the pieces, and what is each one responsible for?→Containers & KubernetesWhy does Kubernetes schedule pods instead of containers? What is a pod actually for?→Containers & KubernetesClusterIP, NodePort, LoadBalancer: what is the difference, and which would you use in production?→Containers & KubernetesWhat does an Ingress do that a Service cannot, and what is an ingress controller?→