Kubernetes Scheduler インテグレーションのメトリクス

Kubernetes Scheduler の監視に使う Datadog の公式インテグレーションです。収集するメトリクスは 43 件あります。

公式の紹介文 (英語)

Monitors the Kubernetes Scheduler

種別の内訳

  • gauge (その時点の値): 38 件
  • count (集計間隔内の発生回数): 5 件

メトリクスの一覧

説明は公式の定義の英文をそのまま載せています。

メトリクス種別単位
kube_scheduler.binding_duration.countNumber of latency in secondsgauge-
kube_scheduler.binding_duration.sumTotal binding latency in secondsgauge-
kube_scheduler.cache.lookupsNumber of equivalence cache lookups, by whether or not a cache entry was foundcount-
kube_scheduler.client.http.requestsNumber of HTTP requests, partitioned by status code, method, and hostcount-
kube_scheduler.client.http.requests_duration.countNumber of client requests. Broken down by verb and URLgauge-
kube_scheduler.client.http.requests_duration.sumTotal latency. Broken down by verb and URLgauge/second
kube_scheduler.gc_duration_seconds.countNumber of the GC invocationgauge-
kube_scheduler.gc_duration_seconds.quantileGC invocation durations quantilesgauge/second
kube_scheduler.gc_duration_seconds.sumGC invocation durations sumgauge/second
kube_scheduler.goroutine_by_scheduling_operationNumber of running goroutines split by the work they do such as binding (alpha; requires k8s v1.26+)gauge-
kube_scheduler.goroutinesNumber of goroutines that currently existgauge-
kube_scheduler.max_fdsMaximum allowed open file descriptorsgauge-
kube_scheduler.open_fdsNumber of open file descriptorsgauge-
kube_scheduler.pending_podsNumber of pending pods, by the queue type (requires k8s v1.15+)gauge-
kube_scheduler.pod_preemption.attemptsNumber of preemption attempts in the cluster till nowcount-
kube_scheduler.pod_preemption.victims.countNumber of selected pods during the latest preemption roundgauge-
kube_scheduler.pod_preemption.victims.sumTotal selected pods during the latest preemption roundgauge-
kube_scheduler.queue.incoming_podsNumber of pods added to scheduling queues by event and queue type (requires k8s v1.17+)count-
kube_scheduler.schedule_attemptsNumber of attempts to schedule pods, by the result. 'unschedulable' means a pod could not be scheduled, while 'error' means an internal scheduler problem.gauge-
kube_scheduler.scheduling.algorithm.predicate_duration.countNumber of scheduling algorithm predicate evaluationgauge-
kube_scheduler.scheduling.algorithm.predicate_duration.sumTotal scheduling algorithm predicate evaluation durationgauge/second
kube_scheduler.scheduling.algorithm.preemption_duration.countNumber of scheduling algorithm preemption evaluationgauge-
kube_scheduler.scheduling.algorithm.preemption_duration.sumTotal scheduling algorithm preemption evaluation durationgauge/second
kube_scheduler.scheduling.algorithm.priority_duration.countNumber of scheduling algorithm priority evaluationgauge-
kube_scheduler.scheduling.algorithm.priority_duration.sumTotal scheduling algorithm priority evaluation durationgauge/second
kube_scheduler.scheduling.algorithm_duration.countNumber of scheduling algorithm latencygauge-
kube_scheduler.scheduling.algorithm_duration.sumTotal scheduling algorithm latencygauge/second
kube_scheduler.scheduling.attempt_duration.countScheduling attempt latency in seconds (scheduling algorithm + binding) (requires k8s v1.23+)gauge-
kube_scheduler.scheduling.attempt_duration.sumTotal scheduling attempt latency in seconds (scheduling algorithm + binding) (requires k8s v1.23+)gauge/second
kube_scheduler.scheduling.e2e_scheduling_duration.countNumber of E2e scheduling latency (scheduling algorithm + binding)gauge-
kube_scheduler.scheduling.e2e_scheduling_duration.sumTotal E2e scheduling latency (scheduling algorithm + binding)gauge/second
kube_scheduler.scheduling.pod.scheduling_attempts.countNumber of attempts to successfully schedule a pod (requires k8s v1.23+)gauge-
kube_scheduler.scheduling.pod.scheduling_attempts.sumTotal number of attempts to successfully schedule a pod (requires k8s v1.23+)gauge-
kube_scheduler.scheduling.pod.scheduling_duration.countE2e latency for a pod being scheduled which may include multiple scheduling attempts (requires k8s v1.23+)gauge-
kube_scheduler.scheduling.pod.scheduling_duration.sumTotal e2e latency for a pod being scheduled which may include multiple scheduling attempts (requires k8s v1.23+)gauge/second
kube_scheduler.scheduling.scheduling_duration.countNumber of scheduling split by sub-parts of the scheduling operationgauge-
kube_scheduler.scheduling.scheduling_duration.quantileScheduling latency quantiles split by sub-parts of the scheduling operationgauge/second
kube_scheduler.scheduling.scheduling_duration.sumTotal scheduling latency split by sub-parts of the scheduling operationgauge/second
kube_scheduler.slis.kubernetes_healthcheckResult of a single scheduler healthcheck (alpha; requires k8s v1.26+)gauge-
kube_scheduler.slis.kubernetes_healthcheck_totalCumulative results of all scheduler healthchecks (alpha; requires k8s v1.26+)count-
kube_scheduler.threadsNumber of OS threads createdgauge-
kube_scheduler.volume_scheduling_duration.countNumber of Volume schedulinggauge-
kube_scheduler.volume_scheduling_duration.sumTotal Volume scheduling stage latencygauge/second

公式ドキュメント

Kubernetes Scheduler インテグレーションの公式ドキュメント

メトリクスの名前と説明は Datadog の公式インテグレーションの定義 (Copyright Datadog, Inc.・BSD-3-Clause) を英文のまま載せています。利用条件の全文は サイトについてにあります。