Kubernetes Onboarding Questionnaire

Q1: Why do we need Prometheus?

Prometheus is used to scrape usage metrics for Kubernetes resources, enabling efficient resource monitoring and cost analysis.

Q2: What metrics do we scrape using Prometheus?

Below are the key metrics we scrape:

CODE

"container_cpu_usage_seconds_total"
	"container_memory_working_set_bytes"
	"kube_pod_container_resource_requests"
	"kube_pod_container_resource_limits"
	"kube_pod_owner"
	"kube_pod_labels"
	"kube_node_status_capacity"
	"kube_node_info"
	"container_fs_writes_bytes_total"
	"container_fs_reads_bytes_total"
	`kube_persistentvolume_created`
	`kube_persistentvolumeclaim_created`
	"kube_persistentvolumeclaim_info"
	"kube_persistentvolume_capacity_bytes"
	"kubelet_volume_stats_used_bytes"
	"kubelet_volume_stats_capacity_bytes"

Q3: How often is Kubernetes cost data refreshed?

Cost data is refreshed once every 24 hours.

Q4: What is Split/Ratio in Kubernetes on Cloud?

Split/Ratio is a method used to distribute the cost of shared Kubernetes resources (e.g., nodes, clusters) among multiple workloads, namespaces, or teams based on defined proportions. Currently, we are following industry-standard ratios and setting the CPU and memory ratio as 87:13 by default. If you would like to update the ratio, you can do so from the K8s Admin page. After applying a split/ratio, you need to wait 24 hours for the cost data to refresh.