Skip to content

Kube Scheduler

By monitoring Kube Scheduler Metrics, it helps configure and optimize the Kube Scheduler, which can improve the resource utilization of the cluster and the performance of applications

Configuration¶

Prerequisites¶

DataKit is installed

Collector Configuration¶

Configure ConfigMap resources in datakit.yaml to collect Kube Scheduler Metrics data

apiVersion: v1
kind: ConfigMap
metadata:
  name: datakit-conf
  namespace: datakit
data:
   kubernetesprometheus.conf: |-  
     [inputs.kubernetesprometheus]
        [[inputs.kubernetesprometheus.instances]]
          role       = "pod"
          namespaces = ["kube-system"]
          selector   = "component=kube_scheduler,tier=control-plane"      
          scrape   = "true"
          scheme   = "https"
          port     = "10259"
          path     = "/metrics"
          interval = "30s"

         [inputs.kubernetesprometheus.instances.custom]
           measurement        = "kube_scheduler"
           job_as_measurement = false
           [inputs.kubernetesprometheus.instances.custom.tags]
             cluster_name_k8s = "k8s-test"           
             node_name        = "__kubernetes_pod_node_name"
             instance         = "__kubernetes_mate_instance"

         [inputs.kubernetesprometheus.instances.auth]
           bearer_token_file = "/var/run/secrets/kubernetes.io/serviceaccount/token"
           [inputs.kubernetesprometheus.instances.auth.tls_config]
             insecure_skip_verify = true
             ca_certs = []
             cert     = ""
             cert_key = ""

Restart datakit

kubectl rollout restart ds  datakit -n datakit

Metrics¶

Kube Scheduler Metric Sets¶

Kube Scheduler Metrics are located under the kube-scheduler metric set. Below is an introduction to the related explanations of Kube Scheduler Metrics.

Metrics	Description	Unit
`scheduler_scheduler_cache_size`	`Number of nodes, Pods, and AssumedPods (pods assumed to be scheduled) in the scheduler cache`	count
`scheduler_pending_pods`	`Number of Pending Pods`	count
`process_cpu_seconds_total`	`Total CPU usage of the kube-scheduler process`	s
`rest_client_request_duration_seconds_bucket`	`Analyze HTTP request latency from the method (Verb) and URL dimensions`	ms
`rest_client_requests_total`	`Analyze HTTP requests from status value (Status Code), method (Method), and host (Host) dimensions`	count
`rest_client_requests_total`	`Total number of HTTP requests initiated by kube-scheduler to kube-apiserver`	count