Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Observe and actbreadcrumb arrow Monitor infrastructurebreadcrumb arrow Kubernetes Monitoringbreadcrumb arrow Configure Kubernetes Monitoringbreadcrumb arrow Manage configuration
Grafana Cloud

Manage your Kubernetes Monitoring configuration

Kubernetes Monitoring gathers metrics, logs, and events, and calculates costs for your infrastructure. It also provides recording rules, alerting rules, and allowlists.

How Kubernetes Monitoring works

Kubernetes Monitoring uses the following to provide its data and visualizations.

cAdvisor

cAdvisor (one per Node) is present on each Node in your Cluster, and emits container resource usage metrics such as CPU usage, memory usage, and disk usage. Alloy collects these metrics and sends them to Grafana Cloud.

Cluster events

Kubernetes Cluster controllers emit information about events concerning the lifecycle of Pods, deployments, and Nodes within the Cluster. Alloy pulls these Cluster events using the Kubernetes API server and converts them into log lines, then sends them to Grafana Cloud logs.

Grafana Alloy

Grafana Alloy:

  • Collects all metrics, Cluster events, and Pod logs
  • Receives traces pushed from applications on Clusters
  • Sends the data to Grafana Cloud

kube-state-metrics

Kubernetes Monitoring uses kube-state-metrics service (one replica, by default) to enable you to see the links between Cluster, Node, Pod, and container. The kube-state-metrics service listens to Kubernetes API server events, and generates Prometheus metrics that document the state of your Cluster’s objects. Over a thousand different metrics provide the status, capacity, and health of individual containers, Pods, deployments, and other resources.

kube-state-metrics:

  • Generates metrics without modification
  • Listens to the Kubernetes API server (not present on each Node)
  • Provides metrics on the state of objects in your Cluster (Pods, Deployments, DaemonSets)

The Kubernetes Monitoring Cluster navigation feature requires the following metrics:

  • kube_namespace_status_phase
  • container_cpu_usage_seconds_total
  • kube_pod_status_phase
  • kube_pod_start_time
  • kube_pod_container_status_restarts_total
  • kube_pod_container_info
  • kube_pod_container_status_waiting_reason
  • kube_daemonset.+
  • kube_replicaset.+
  • kube_statefulset.+
  • kube_job.+
  • kube_node.+
  • kube_cluster.+
  • node_cpu_seconds_total
  • node_memory_MemAvailable_bytes
  • node_filesystem_size_bytes
  • node_namespace_pod_container
  • container_memory_working_set_bytes

kubelet

kubelet (one per Node):

  • Is the primary Node agent present on each Node in the Cluster
  • Emits metrics specific to the kubelet process like kubelet_running_pods and kubelet_running_container_count
  • Ensures containers are running
  • Provides metrics on Pods and their containers

Grafana Alloy collects these metrics and sends them to Grafana Cloud.

Kubernetes mixins

Kubernetes Monitoring is heavily indebted to the open source kubernetes-mixin project, from which the recording and alerting rules are derived. Grafana Labs continue to contribute bug fixes and new features upstream.

Node Exporter

The Prometheus exporter node-exporter runs as a DaemonSet on the Cluster to:

  • Gather metrics on hardware and OS for Linux Nodes in the Cluster
  • Emit Prometheus metrics for the health and state of the Nodes in your Cluster

Grafana Alloy collects these metrics and sends them to Grafana Cloud.

OpenCost

Kubernetes Monitoring uses the combination of OpenCost and Grafana to allow you to monitor and manage costs related to your Kubernetes Cluster. For more details, refer to Manage costs.

Node logs

These are logs written by host processes, not by anything running inside a container. Some Kubernetes nodes run system services as systemd units: kubelet (schedules and manages pods), containerd or docker (the container runtime), and kernel/network subsystems. When these components log errors, crashes, or state changes, those entries go to journald on the host.

Pod logs

Alloy pulls Pod logs from the workloads running within containers, and sends them to Loki.

Note

Log entries must be sent to a Loki data source with cluster, namespace, and pod labels.

Traces

Traces generated by applications within the Cluster are pushed to Grafana Alloy. The address options listed during the process of configuring with the Helm chart contain the configuration endpoints where traces can be pushed.

Windows Exporter

When monitoring Windows Nodes, the configuration installs the windows-exporter DaemonSet to ensure metrics are available for scraping.

Recording rules

Recording rules calculate in advance any mathematical expressions that are frequently needed or take a lot of computation, and then save the result as a new set of time series. This decreases the time for query and calculation.

Note

Recording rules may emit time series with the same metric name, but different labels. To modify these programmatically, refer to Set up Alerting for Cloud.

Kubernetes Monitoring includes the following recording rules to speed up queries and the evaluation of alerting rules.

apiserver_request:availability30d
apiserver_request:availability30d
apiserver_request:availability30d
apiserver_request:burnrate1d
apiserver_request:burnrate1d
apiserver_request:burnrate1h
apiserver_request:burnrate1h
apiserver_request:burnrate2h
apiserver_request:burnrate2h
apiserver_request:burnrate30m
apiserver_request:burnrate30m
apiserver_request:burnrate3d
apiserver_request:burnrate3d
apiserver_request:burnrate5m
apiserver_request:burnrate5m
apiserver_request:burnrate6h
apiserver_request:burnrate6h

cluster:namespace:pod_cpu:active:kube_pod_container_resource_limits cluster:namespace:pod_cpu:active:kube_pod_container_resource_requests cluster:namespace:pod_memory:active:kube_pod_container_resource_limits cluster:namespace:pod_memory:active:kube_pod_container_resource_requests

cluster_quantile:apiserver_request_sli_duration_seconds:histogram_quantile cluster_quantile:apiserver_request_sli_duration_seconds:histogram_quantile cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile

cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase1h cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase1h cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d

code:apiserver_request_total:increase30d
code:apiserver_request_total:increase30d

code_resource:apiserver_request_total:rate5m
code_resource:apiserver_request_total:rate5m

code_verb:apiserver_request_total:increase1h
code_verb:apiserver_request_total:increase1h
code_verb:apiserver_request_total:increase1h
code_verb:apiserver_request_total:increase1h
code_verb:apiserver_request_total:increase30d

container_cpu_usage_seconds_total
container_memory_rss
container_memory_working_set_bytes

namespace_cpu:kube_pod_container_resource_limits:sum namespace_cpu:kube_pod_container_resource_requests:sum namespace_memory:kube_pod_container_resource_limits:sum namespace_memory:kube_pod_container_resource_requests:sum namespace_workload_pod:kube_pod_owner:relabel
namespace_workload_pod:kube_pod_owner:relabel
namespace_workload_pod:kube_pod_owner:relabel
namespace_workload_pod:kube_pod_owner:relabel

node_namespace_pod_container:container_cpu_usage_seconds_total:sum_irate node_namespace_pod_container:container_memory_cache node_namespace_pod_container:container_memory_rss node_namespace_pod_container:container_memory_swap node_namespace_pod_container:container_memory_working_set_bytes

node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile

Alerting rules

Kubernetes Monitoring comes with preconfigured alerting rules to alert on conditions such as “Pods crash looping” and “Pods getting stuck in not ready”. The following alerting rules create alerts to notify you when issues arise with your Clusters and their workloads.

To learn more, refer to the upstream Kubernetes-Mixin’s Kubernetes Alert Runbooks page. To update programmatically the alerting rule links to point your own runbooks in these preconfigured alerts, use a tool like cortex-tools or grizzly.

kubelet alerting rules

  • KubeNodeNotReady
  • KubeNodeReadinessFlapping
  • KubeNodeUnreachable
  • KubeletClientCertificateExpiration - 7 day expiration
  • KubeletClientCertificateExpiration - 1 day expiration
  • KubeletDown
  • KubeletPlegDurationHigh
  • KubeletPodStartUpLatencyHigh
  • KubeletServerCertificateExpiration - 7 day expiration
  • KubeletServerCertificateExpiration - 1 day expiration
  • KubeletClientCertificateRenewalErrors
  • KubeletServerCertificateRenewalErrors
  • KubeletTooManyPods

Kubernetes alerting rules

  • KubeContainerWaiting
  • KubeDaemonSetMisScheduled
  • KubeDaemonSetNotScheduled
  • KubeDaemonSetRolloutStuck
  • KubeDeploymentGenerationMismatch
  • KubeDeploymentReplicasMismatch
  • KubeDeploymentRolloutStuck
  • KubeHpaMaxedOut
  • KubeHpaReplicasMismatch
  • KubeJobFailed
  • KubeJobNotCompleted
  • KubePodCrashLooping
  • KubePodNotReady
  • KubeStatefulSetGenerationMismatch
  • KubeStatefulSetReplicasMismatch
  • KubeStatefulSetUpdateNotRolledOut

Kubernetes API alerting rules

  • KubeAggregatedAPIDown
  • KubeAggregatedAPIErrors
  • KubeAPIDown
  • KubeAPIErrorBudgetBurn
  • KubeAPIErrorBudgetBurn
  • KubeAPIErrorBudgetBurn
  • KubeAPIErrorBudgetBurn
  • KubeAPITerminatedRequests
  • KubeClientCertificateExpiration - less than 7 days
  • KubeClientCertificateExpiration - less than 1 day

Kubernetes resource usage alerting rules

  • CPUThrottlingHigh
  • KubeCPUOvercommit
  • KubeCPUQuotaOvercommit
  • KubeMemoryOvercommit
  • KubeMemoryQuotaOvercommit
  • KubeQuotaAlmostFull
  • KubeQuotaExceeded
  • KubeQuotaFullyUsed

Kubernetes storage alerting rules

  • KubePersistentVolumeErrors
  • KubePersistentVolumeFillingUp - 1 hour
  • KubePersistentVolumeFillingUp - 1 minute
  • KubePersistentVolumeInodesFillingUp - 1 hour
  • KubePersistentVolumeInodesFillingUp - 1 minute

Kubernetes system alerting rules

  • KubeClientErrors
  • KubeVersionMismatch

Metrics management and control

If your account is based on billable series, there are ways to control and manage metrics.

Identify unnecessary or duplicate metrics

To identify unnecessary or duplicate metrics generated from a Cluster, you can:

  • Use the Cardinality page to discover on a Cluster-by-Cluster basis where all your active series are coming from. From the main menu, click Configuration and then the Cardinality tab.
    Cardinality page within the app
    Cardinality page within the app
  • Analyze current metrics usage and associated costs from the billing and usage dashboard located in your Grafana instance.

Analyze usage

For techniques to analyze usage, refer to Analyze Prometheus metrics costs.

Reduce usage

Use and refine an allowlist to reduce metrics to only those you want to receive. An allowlist is a set of metrics and labels that you want to gather while all others are dropped. Most metric sources apply a default allowlist that keeps the minimal set of metrics Kubernetes Monitoring needs, because metricsTuning.useDefaultAllowList defaults to true. The control plane sources are the exception: apiServer, kubeControllerManager, kubeDNS, kubeProxy, and kubeScheduler have no default allowlist and keep everything they scrape.

To adjust a source, set metricsTuning.includeMetrics to keep additional metrics and metricsTuning.excludeMetrics to drop metrics you don’t need. Both accept regular expressions. On a source that has a default allowlist, set metricsTuning.useDefaultAllowList to false to stop filtering it.

When keeping and dropping aren’t enough, use extraMetricProcessingRules, which takes Alloy relabeling rules. For the keep and drop actions and the other fields a rule accepts, refer to rule block.

Relabeling works the same way it does in Prometheus, so Relabeling rule fields is useful background on what each field does. The examples there use Prometheus YAML, so write the rules themselves in Alloy syntax.

These keys belong to the individual metric source, under the feature that collects it, such as clusterMetrics.kubelet or hostMetrics.linuxHosts. They’re configured locally on your Cluster rather than from Grafana Cloud. For examples, refer to Processing and labeling and Metrics tuning in the Helm chart repository.

To learn more about how dropping metrics affects cost, refer to Relabel Prometheus metrics to reduce usage.

Billable series

If your account does not use host-hours pricing, pricing is based on billable series. To learn more about this pricing model, refer to Active series and DPM.

Default telemetry data collection (also called active series) varies depending on your Kubernetes Cluster size (number of Nodes) and running workloads (number of Pods, containers, Deployments, etc.).

Active series usage scales with both. Each additional Node adds series from the collectors and exporters that run on every Node, and each additional Pod or container adds series describing Cluster state. Targets beyond the defaults, such as application metrics, API server metrics, and scheduler metrics, add more series on top.

Because usage also varies with your Cloud provider and Kubernetes version, measure your own Cluster rather than working from a general estimate. To see where your active series come from, use the Cardinality page described in Metrics management and control.

Logs management

Control and manage logs by:

  • Only collecting logs from Pods in certain namespaces
  • Dropping logs based on content

For more on analyzing, customizing, and deduplicating logs, refer to Logs in Explore.

Limit logs to only Pods from certain namespaces

By default, Kubernetes Monitoring gathers logs from Pods in every namespace. However, you may only want or need logs from Pods in certain namespaces.

Set the namespace list in your values file. On the Pod logs feature you use, such as podLogsViaLoki, the namespaces key limits collection to the namespaces you list, and excludeNamespaces drops the namespaces you list.

For example, the following gathers logs only from the production and staging namespaces:

YAML 
podLogsViaLoki:
  namespaces:
    - production
    - staging

When namespace and label matching aren’t enough, set extraDiscoveryRules on the same feature, which takes Alloy relabeling rules. Refer to Discovery.

Drop logs based on content

Similarly to filtering to specific namespaces, you can use Loki processing rules to further process and optionally drop log lines. Set them with extraLogProcessingStages on the Pod logs feature you use.

For example, this processing stage drops any log lines that contain the word debug:

YAML 
podLogsViaLoki:
  extraLogProcessingStages: |-
    stage.drop {
      expression  = ".*(debug|DEBUG).*"
    }

Manage Assistant preferences

The Manage app tab on the Kubernetes Monitoring Configuration page includes preferences for Grafana Assistant.

Auto-run assistant health checks

The Auto-run assistant health checks toggle controls whether Assistant health checks run automatically on every Kubernetes detail page.

  • When the toggle is on, Grafana Assistant executes Prometheus and Loki queries in the background on every Cluster, namespace, workload, Pod, container, and Node detail page, then displays the health check message.
  • When the toggle is off, the check doesn’t run automatically. Refresh it manually from the banner, or open the Assistant from a panel to ask about a page.

Turn off the auto health check to reduce background query load and Assistant usage.

The toggle is saved in your browser’s local storage, so it’s a per-user, per-browser preference that doesn’t affect teammates or other browsers.

**Auto-run assistant health checks** toggle on the **Manage app** tab
Auto-run assistant health checks toggle on the Manage app tab

Deactivate Kubernetes Monitoring

To deactivate Kubernetes Monitoring, refer to Deactivate Kubernetes Monitoring.

Get support

To open a support ticket, navigate to your Grafana Cloud Portal, and click Open a Support Ticket.