1243 lines
88 KiB
Plaintext
1243 lines
88 KiB
Plaintext
{
|
|
"apiVersion": "monitoring.coreos.com/v1",
|
|
"kind": "PrometheusRule",
|
|
"metadata": {
|
|
"labels": {
|
|
"app.kubernetes.io/name": "kube-prometheus",
|
|
"app.kubernetes.io/part-of": "kube-prometheus",
|
|
"prometheus": "k8s",
|
|
"role": "alert-rules"
|
|
},
|
|
"name": "kubernetes-monitoring-rules",
|
|
"namespace": "monitoring"
|
|
},
|
|
"spec": {
|
|
"groups": [
|
|
{
|
|
"name": "kubernetes-apps",
|
|
"rules": [
|
|
{
|
|
"alert": "KubePodCrashLooping",
|
|
"annotations": {
|
|
"description": "Pod {{ $labels.namespace }}/{{ $labels.pod }} ({{ $labels.container }}) is in waiting state (reason: \"CrashLoopBackOff\").",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepodcrashlooping",
|
|
"summary": "Pod is crash looping."
|
|
},
|
|
"expr": "max_over_time(kube_pod_container_status_waiting_reason{reason=\"CrashLoopBackOff\", job=\"kube-state-metrics\"}[5m]) >= 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubePodNotReady",
|
|
"annotations": {
|
|
"description": "Pod {{ $labels.namespace }}/{{ $labels.pod }} has been in a non-ready state for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepodnotready",
|
|
"summary": "Pod has been in a non-ready state for more than 15 minutes."
|
|
},
|
|
"expr": "sum by (namespace, pod, cluster) (\n max by(namespace, pod, cluster) (\n kube_pod_status_phase{job=\"kube-state-metrics\", phase=~\"Pending|Unknown|Failed\"}\n ) * on(namespace, pod, cluster) group_left(owner_kind) topk by(namespace, pod, cluster) (\n 1, max by(namespace, pod, owner_kind, cluster) (kube_pod_owner{owner_kind!=\"Job\"})\n )\n) > 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDeploymentGenerationMismatch",
|
|
"annotations": {
|
|
"description": "Deployment generation for {{ $labels.namespace }}/{{ $labels.deployment }} does not match, this indicates that the Deployment has failed but has not been rolled back.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedeploymentgenerationmismatch",
|
|
"summary": "Deployment generation mismatch due to possible roll-back"
|
|
},
|
|
"expr": "kube_deployment_status_observed_generation{job=\"kube-state-metrics\"}\n !=\nkube_deployment_metadata_generation{job=\"kube-state-metrics\"}\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDeploymentReplicasMismatch",
|
|
"annotations": {
|
|
"description": "Deployment {{ $labels.namespace }}/{{ $labels.deployment }} has not matched the expected number of replicas for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedeploymentreplicasmismatch",
|
|
"summary": "Deployment has not matched the expected number of replicas."
|
|
},
|
|
"expr": "(\n kube_deployment_spec_replicas{job=\"kube-state-metrics\"}\n >\n kube_deployment_status_replicas_available{job=\"kube-state-metrics\"}\n) and (\n changes(kube_deployment_status_replicas_updated{job=\"kube-state-metrics\"}[10m])\n ==\n 0\n)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDeploymentRolloutStuck",
|
|
"annotations": {
|
|
"description": "Rollout of deployment {{ $labels.namespace }}/{{ $labels.deployment }} is not progressing for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedeploymentrolloutstuck",
|
|
"summary": "Deployment rollout is not progressing."
|
|
},
|
|
"expr": "kube_deployment_status_condition{condition=\"Progressing\", status=\"false\",job=\"kube-state-metrics\"}\n!= 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeStatefulSetReplicasMismatch",
|
|
"annotations": {
|
|
"description": "StatefulSet {{ $labels.namespace }}/{{ $labels.statefulset }} has not matched the expected number of replicas for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubestatefulsetreplicasmismatch",
|
|
"summary": "StatefulSet has not matched the expected number of replicas."
|
|
},
|
|
"expr": "(\n kube_statefulset_status_replicas_ready{job=\"kube-state-metrics\"}\n !=\n kube_statefulset_status_replicas{job=\"kube-state-metrics\"}\n) and (\n changes(kube_statefulset_status_replicas_updated{job=\"kube-state-metrics\"}[10m])\n ==\n 0\n)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeStatefulSetGenerationMismatch",
|
|
"annotations": {
|
|
"description": "StatefulSet generation for {{ $labels.namespace }}/{{ $labels.statefulset }} does not match, this indicates that the StatefulSet has failed but has not been rolled back.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubestatefulsetgenerationmismatch",
|
|
"summary": "StatefulSet generation mismatch due to possible roll-back"
|
|
},
|
|
"expr": "kube_statefulset_status_observed_generation{job=\"kube-state-metrics\"}\n !=\nkube_statefulset_metadata_generation{job=\"kube-state-metrics\"}\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeStatefulSetUpdateNotRolledOut",
|
|
"annotations": {
|
|
"description": "StatefulSet {{ $labels.namespace }}/{{ $labels.statefulset }} update has not been rolled out.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubestatefulsetupdatenotrolledout",
|
|
"summary": "StatefulSet update has not been rolled out."
|
|
},
|
|
"expr": "(\n max by(namespace, statefulset) (\n kube_statefulset_status_current_revision{job=\"kube-state-metrics\"}\n unless\n kube_statefulset_status_update_revision{job=\"kube-state-metrics\"}\n )\n *\n (\n kube_statefulset_replicas{job=\"kube-state-metrics\"}\n !=\n kube_statefulset_status_replicas_updated{job=\"kube-state-metrics\"}\n )\n) and (\n changes(kube_statefulset_status_replicas_updated{job=\"kube-state-metrics\"}[5m])\n ==\n 0\n)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDaemonSetRolloutStuck",
|
|
"annotations": {
|
|
"description": "DaemonSet {{ $labels.namespace }}/{{ $labels.daemonset }} has not finished or progressed for at least 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedaemonsetrolloutstuck",
|
|
"summary": "DaemonSet rollout is stuck."
|
|
},
|
|
"expr": "(\n (\n kube_daemonset_status_current_number_scheduled{job=\"kube-state-metrics\"}\n !=\n kube_daemonset_status_desired_number_scheduled{job=\"kube-state-metrics\"}\n ) or (\n kube_daemonset_status_number_misscheduled{job=\"kube-state-metrics\"}\n !=\n 0\n ) or (\n kube_daemonset_status_updated_number_scheduled{job=\"kube-state-metrics\"}\n !=\n kube_daemonset_status_desired_number_scheduled{job=\"kube-state-metrics\"}\n ) or (\n kube_daemonset_status_number_available{job=\"kube-state-metrics\"}\n !=\n kube_daemonset_status_desired_number_scheduled{job=\"kube-state-metrics\"}\n )\n) and (\n changes(kube_daemonset_status_updated_number_scheduled{job=\"kube-state-metrics\"}[5m])\n ==\n 0\n)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeContainerWaiting",
|
|
"annotations": {
|
|
"description": "pod/{{ $labels.pod }} in namespace {{ $labels.namespace }} on container {{ $labels.container}} has been in waiting state for longer than 1 hour.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubecontainerwaiting",
|
|
"summary": "Pod container waiting longer than 1 hour"
|
|
},
|
|
"expr": "sum by (namespace, pod, container, cluster) (kube_pod_container_status_waiting_reason{job=\"kube-state-metrics\"}) > 0\n",
|
|
"for": "1h",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDaemonSetNotScheduled",
|
|
"annotations": {
|
|
"description": "{{ $value }} Pods of DaemonSet {{ $labels.namespace }}/{{ $labels.daemonset }} are not scheduled.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedaemonsetnotscheduled",
|
|
"summary": "DaemonSet pods are not scheduled."
|
|
},
|
|
"expr": "kube_daemonset_status_desired_number_scheduled{job=\"kube-state-metrics\"}\n -\nkube_daemonset_status_current_number_scheduled{job=\"kube-state-metrics\"} > 0\n",
|
|
"for": "10m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeDaemonSetMisScheduled",
|
|
"annotations": {
|
|
"description": "{{ $value }} Pods of DaemonSet {{ $labels.namespace }}/{{ $labels.daemonset }} are running where they are not supposed to run.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubedaemonsetmisscheduled",
|
|
"summary": "DaemonSet pods are misscheduled."
|
|
},
|
|
"expr": "kube_daemonset_status_number_misscheduled{job=\"kube-state-metrics\"} > 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeJobNotCompleted",
|
|
"annotations": {
|
|
"description": "Job {{ $labels.namespace }}/{{ $labels.job_name }} is taking more than {{ \"43200\" | humanizeDuration }} to complete.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubejobnotcompleted",
|
|
"summary": "Job did not complete in time"
|
|
},
|
|
"expr": "time() - max by(namespace, job_name, cluster) (kube_job_status_start_time{job=\"kube-state-metrics\"}\n and\nkube_job_status_active{job=\"kube-state-metrics\"} > 0) > 43200\n",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeJobFailed",
|
|
"annotations": {
|
|
"description": "Job {{ $labels.namespace }}/{{ $labels.job_name }} failed to complete. Removing failed job after investigation should clear this alert.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubejobfailed",
|
|
"summary": "Job failed to complete."
|
|
},
|
|
"expr": "kube_job_failed{job=\"kube-state-metrics\"} > 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeHpaReplicasMismatch",
|
|
"annotations": {
|
|
"description": "HPA {{ $labels.namespace }}/{{ $labels.horizontalpodautoscaler }} has not matched the desired number of replicas for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubehpareplicasmismatch",
|
|
"summary": "HPA has not matched desired number of replicas."
|
|
},
|
|
"expr": "(kube_horizontalpodautoscaler_status_desired_replicas{job=\"kube-state-metrics\"}\n !=\nkube_horizontalpodautoscaler_status_current_replicas{job=\"kube-state-metrics\"})\n and\n(kube_horizontalpodautoscaler_status_current_replicas{job=\"kube-state-metrics\"}\n >\nkube_horizontalpodautoscaler_spec_min_replicas{job=\"kube-state-metrics\"})\n and\n(kube_horizontalpodautoscaler_status_current_replicas{job=\"kube-state-metrics\"}\n <\nkube_horizontalpodautoscaler_spec_max_replicas{job=\"kube-state-metrics\"})\n and\nchanges(kube_horizontalpodautoscaler_status_current_replicas{job=\"kube-state-metrics\"}[15m]) == 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeHpaMaxedOut",
|
|
"annotations": {
|
|
"description": "HPA {{ $labels.namespace }}/{{ $labels.horizontalpodautoscaler }} has been running at max replicas for longer than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubehpamaxedout",
|
|
"summary": "HPA is running at max replicas"
|
|
},
|
|
"expr": "kube_horizontalpodautoscaler_status_current_replicas{job=\"kube-state-metrics\"}\n ==\nkube_horizontalpodautoscaler_spec_max_replicas{job=\"kube-state-metrics\"}\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-resources",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeCPUOvercommit",
|
|
"annotations": {
|
|
"description": "Cluster {{ $labels.cluster }} has overcommitted CPU resource requests for Pods by {{ $value }} CPU shares and cannot tolerate node failure.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubecpuovercommit",
|
|
"summary": "Cluster has overcommitted CPU resource requests."
|
|
},
|
|
"expr": "sum(namespace_cpu:kube_pod_container_resource_requests:sum{}) by (cluster) - (sum(kube_node_status_allocatable{job=\"kube-state-metrics\",resource=\"cpu\"}) by (cluster) - max(kube_node_status_allocatable{job=\"kube-state-metrics\",resource=\"cpu\"}) by (cluster)) > 0\nand\n(sum(kube_node_status_allocatable{job=\"kube-state-metrics\",resource=\"cpu\"}) by (cluster) - max(kube_node_status_allocatable{job=\"kube-state-metrics\",resource=\"cpu\"}) by (cluster)) > 0\n",
|
|
"for": "10m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeMemoryOvercommit",
|
|
"annotations": {
|
|
"description": "Cluster {{ $labels.cluster }} has overcommitted memory resource requests for Pods by {{ $value | humanize }} bytes and cannot tolerate node failure.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubememoryovercommit",
|
|
"summary": "Cluster has overcommitted memory resource requests."
|
|
},
|
|
"expr": "sum(namespace_memory:kube_pod_container_resource_requests:sum{}) by (cluster) - (sum(kube_node_status_allocatable{resource=\"memory\", job=\"kube-state-metrics\"}) by (cluster) - max(kube_node_status_allocatable{resource=\"memory\", job=\"kube-state-metrics\"}) by (cluster)) > 0\nand\n(sum(kube_node_status_allocatable{resource=\"memory\", job=\"kube-state-metrics\"}) by (cluster) - max(kube_node_status_allocatable{resource=\"memory\", job=\"kube-state-metrics\"}) by (cluster)) > 0\n",
|
|
"for": "10m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeCPUQuotaOvercommit",
|
|
"annotations": {
|
|
"description": "Cluster {{ $labels.cluster }} has overcommitted CPU resource requests for Namespaces.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubecpuquotaovercommit",
|
|
"summary": "Cluster has overcommitted CPU resource requests."
|
|
},
|
|
"expr": "sum(min without(resource) (kube_resourcequota{job=\"kube-state-metrics\", type=\"hard\", resource=~\"(cpu|requests.cpu)\"})) by (cluster)\n /\nsum(kube_node_status_allocatable{resource=\"cpu\", job=\"kube-state-metrics\"}) by (cluster)\n > 1.5\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeMemoryQuotaOvercommit",
|
|
"annotations": {
|
|
"description": "Cluster {{ $labels.cluster }} has overcommitted memory resource requests for Namespaces.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubememoryquotaovercommit",
|
|
"summary": "Cluster has overcommitted memory resource requests."
|
|
},
|
|
"expr": "sum(min without(resource) (kube_resourcequota{job=\"kube-state-metrics\", type=\"hard\", resource=~\"(memory|requests.memory)\"})) by (cluster)\n /\nsum(kube_node_status_allocatable{resource=\"memory\", job=\"kube-state-metrics\"}) by (cluster)\n > 1.5\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeQuotaAlmostFull",
|
|
"annotations": {
|
|
"description": "Namespace {{ $labels.namespace }} is using {{ $value | humanizePercentage }} of its {{ $labels.resource }} quota.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubequotaalmostfull",
|
|
"summary": "Namespace quota is going to be full."
|
|
},
|
|
"expr": "kube_resourcequota{job=\"kube-state-metrics\", type=\"used\"}\n / ignoring(instance, job, type)\n(kube_resourcequota{job=\"kube-state-metrics\", type=\"hard\"} > 0)\n > 0.9 < 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "info"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeQuotaFullyUsed",
|
|
"annotations": {
|
|
"description": "Namespace {{ $labels.namespace }} is using {{ $value | humanizePercentage }} of its {{ $labels.resource }} quota.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubequotafullyused",
|
|
"summary": "Namespace quota is fully used."
|
|
},
|
|
"expr": "kube_resourcequota{job=\"kube-state-metrics\", type=\"used\"}\n / ignoring(instance, job, type)\n(kube_resourcequota{job=\"kube-state-metrics\", type=\"hard\"} > 0)\n == 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "info"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeQuotaExceeded",
|
|
"annotations": {
|
|
"description": "Namespace {{ $labels.namespace }} is using {{ $value | humanizePercentage }} of its {{ $labels.resource }} quota.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubequotaexceeded",
|
|
"summary": "Namespace quota has exceeded the limits."
|
|
},
|
|
"expr": "kube_resourcequota{job=\"kube-state-metrics\", type=\"used\"}\n / ignoring(instance, job, type)\n(kube_resourcequota{job=\"kube-state-metrics\", type=\"hard\"} > 0)\n > 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "CPUThrottlingHigh",
|
|
"annotations": {
|
|
"description": "{{ $value | humanizePercentage }} throttling of CPU in namespace {{ $labels.namespace }} for container {{ $labels.container }} in pod {{ $labels.pod }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/cputhrottlinghigh",
|
|
"summary": "Processes experience elevated CPU throttling."
|
|
},
|
|
"expr": "sum(increase(container_cpu_cfs_throttled_periods_total{container!=\"\", }[5m])) by (cluster, container, pod, namespace)\n /\nsum(increase(container_cpu_cfs_periods_total{}[5m])) by (cluster, container, pod, namespace)\n > ( 25 / 100 )\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "info"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-storage",
|
|
"rules": [
|
|
{
|
|
"alert": "KubePersistentVolumeFillingUp",
|
|
"annotations": {
|
|
"description": "The PersistentVolume claimed by {{ $labels.persistentvolumeclaim }} in Namespace {{ $labels.namespace }} {{ with $labels.cluster -}} on Cluster {{ . }} {{- end }} is only {{ $value | humanizePercentage }} free.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepersistentvolumefillingup",
|
|
"summary": "PersistentVolume is filling up."
|
|
},
|
|
"expr": "(\n kubelet_volume_stats_available_bytes{job=\"kubelet\", metrics_path=\"/metrics\"}\n /\n kubelet_volume_stats_capacity_bytes{job=\"kubelet\", metrics_path=\"/metrics\"}\n) < 0.03\nand\nkubelet_volume_stats_used_bytes{job=\"kubelet\", metrics_path=\"/metrics\"} > 0\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_access_mode{ access_mode=\"ReadOnlyMany\"} == 1\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_labels{label_excluded_from_alerts=\"true\"} == 1\n",
|
|
"for": "1m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubePersistentVolumeFillingUp",
|
|
"annotations": {
|
|
"description": "Based on recent sampling, the PersistentVolume claimed by {{ $labels.persistentvolumeclaim }} in Namespace {{ $labels.namespace }} {{ with $labels.cluster -}} on Cluster {{ . }} {{- end }} is expected to fill up within four days. Currently {{ $value | humanizePercentage }} is available.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepersistentvolumefillingup",
|
|
"summary": "PersistentVolume is filling up."
|
|
},
|
|
"expr": "(\n kubelet_volume_stats_available_bytes{job=\"kubelet\", metrics_path=\"/metrics\"}\n /\n kubelet_volume_stats_capacity_bytes{job=\"kubelet\", metrics_path=\"/metrics\"}\n) < 0.15\nand\nkubelet_volume_stats_used_bytes{job=\"kubelet\", metrics_path=\"/metrics\"} > 0\nand\npredict_linear(kubelet_volume_stats_available_bytes{job=\"kubelet\", metrics_path=\"/metrics\"}[6h], 4 * 24 * 3600) < 0\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_access_mode{ access_mode=\"ReadOnlyMany\"} == 1\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_labels{label_excluded_from_alerts=\"true\"} == 1\n",
|
|
"for": "1h",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubePersistentVolumeInodesFillingUp",
|
|
"annotations": {
|
|
"description": "The PersistentVolume claimed by {{ $labels.persistentvolumeclaim }} in Namespace {{ $labels.namespace }} {{ with $labels.cluster -}} on Cluster {{ . }} {{- end }} only has {{ $value | humanizePercentage }} free inodes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepersistentvolumeinodesfillingup",
|
|
"summary": "PersistentVolumeInodes are filling up."
|
|
},
|
|
"expr": "(\n kubelet_volume_stats_inodes_free{job=\"kubelet\", metrics_path=\"/metrics\"}\n /\n kubelet_volume_stats_inodes{job=\"kubelet\", metrics_path=\"/metrics\"}\n) < 0.03\nand\nkubelet_volume_stats_inodes_used{job=\"kubelet\", metrics_path=\"/metrics\"} > 0\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_access_mode{ access_mode=\"ReadOnlyMany\"} == 1\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_labels{label_excluded_from_alerts=\"true\"} == 1\n",
|
|
"for": "1m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubePersistentVolumeInodesFillingUp",
|
|
"annotations": {
|
|
"description": "Based on recent sampling, the PersistentVolume claimed by {{ $labels.persistentvolumeclaim }} in Namespace {{ $labels.namespace }} {{ with $labels.cluster -}} on Cluster {{ . }} {{- end }} is expected to run out of inodes within four days. Currently {{ $value | humanizePercentage }} of its inodes are free.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepersistentvolumeinodesfillingup",
|
|
"summary": "PersistentVolumeInodes are filling up."
|
|
},
|
|
"expr": "(\n kubelet_volume_stats_inodes_free{job=\"kubelet\", metrics_path=\"/metrics\"}\n /\n kubelet_volume_stats_inodes{job=\"kubelet\", metrics_path=\"/metrics\"}\n) < 0.15\nand\nkubelet_volume_stats_inodes_used{job=\"kubelet\", metrics_path=\"/metrics\"} > 0\nand\npredict_linear(kubelet_volume_stats_inodes_free{job=\"kubelet\", metrics_path=\"/metrics\"}[6h], 4 * 24 * 3600) < 0\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_access_mode{ access_mode=\"ReadOnlyMany\"} == 1\nunless on(cluster, namespace, persistentvolumeclaim)\nkube_persistentvolumeclaim_labels{label_excluded_from_alerts=\"true\"} == 1\n",
|
|
"for": "1h",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubePersistentVolumeErrors",
|
|
"annotations": {
|
|
"description": "The persistent volume {{ $labels.persistentvolume }} {{ with $labels.cluster -}} on Cluster {{ . }} {{- end }} has status {{ $labels.phase }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubepersistentvolumeerrors",
|
|
"summary": "PersistentVolume is having issues with provisioning."
|
|
},
|
|
"expr": "kube_persistentvolume_status_phase{phase=~\"Failed|Pending\",job=\"kube-state-metrics\"} > 0\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-system",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeVersionMismatch",
|
|
"annotations": {
|
|
"description": "There are {{ $value }} different semantic versions of Kubernetes components running.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeversionmismatch",
|
|
"summary": "Different semantic versions of Kubernetes components running."
|
|
},
|
|
"expr": "count by (cluster) (count by (git_version, cluster) (label_replace(kubernetes_build_info{job!~\"kube-dns|coredns\"},\"git_version\",\"$1\",\"git_version\",\"(v[0-9]*.[0-9]*).*\"))) > 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeClientErrors",
|
|
"annotations": {
|
|
"description": "Kubernetes API server client '{{ $labels.job }}/{{ $labels.instance }}' is experiencing {{ $value | humanizePercentage }} errors.'",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeclienterrors",
|
|
"summary": "Kubernetes API server client is experiencing errors."
|
|
},
|
|
"expr": "(sum(rate(rest_client_requests_total{job=\"apiserver\",code=~\"5..\"}[5m])) by (cluster, instance, job, namespace)\n /\nsum(rate(rest_client_requests_total{job=\"apiserver\"}[5m])) by (cluster, instance, job, namespace))\n> 0.01\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kube-apiserver-slos",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeAPIErrorBudgetBurn",
|
|
"annotations": {
|
|
"description": "The API server is burning too much error budget.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapierrorbudgetburn",
|
|
"summary": "The API server is burning too much error budget."
|
|
},
|
|
"expr": "sum(apiserver_request:burnrate1h) > (14.40 * 0.01000)\nand\nsum(apiserver_request:burnrate5m) > (14.40 * 0.01000)\n",
|
|
"for": "2m",
|
|
"labels": {
|
|
"long": "1h",
|
|
"severity": "critical",
|
|
"short": "5m"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAPIErrorBudgetBurn",
|
|
"annotations": {
|
|
"description": "The API server is burning too much error budget.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapierrorbudgetburn",
|
|
"summary": "The API server is burning too much error budget."
|
|
},
|
|
"expr": "sum(apiserver_request:burnrate6h) > (6.00 * 0.01000)\nand\nsum(apiserver_request:burnrate30m) > (6.00 * 0.01000)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"long": "6h",
|
|
"severity": "critical",
|
|
"short": "30m"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAPIErrorBudgetBurn",
|
|
"annotations": {
|
|
"description": "The API server is burning too much error budget.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapierrorbudgetburn",
|
|
"summary": "The API server is burning too much error budget."
|
|
},
|
|
"expr": "sum(apiserver_request:burnrate1d) > (3.00 * 0.01000)\nand\nsum(apiserver_request:burnrate2h) > (3.00 * 0.01000)\n",
|
|
"for": "1h",
|
|
"labels": {
|
|
"long": "1d",
|
|
"severity": "warning",
|
|
"short": "2h"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAPIErrorBudgetBurn",
|
|
"annotations": {
|
|
"description": "The API server is burning too much error budget.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapierrorbudgetburn",
|
|
"summary": "The API server is burning too much error budget."
|
|
},
|
|
"expr": "sum(apiserver_request:burnrate3d) > (1.00 * 0.01000)\nand\nsum(apiserver_request:burnrate6h) > (1.00 * 0.01000)\n",
|
|
"for": "3h",
|
|
"labels": {
|
|
"long": "3d",
|
|
"severity": "warning",
|
|
"short": "6h"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-system-apiserver",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeClientCertificateExpiration",
|
|
"annotations": {
|
|
"description": "A client certificate used to authenticate to kubernetes apiserver is expiring in less than 7.0 days.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeclientcertificateexpiration",
|
|
"summary": "Client certificate is about to expire."
|
|
},
|
|
"expr": "apiserver_client_certificate_expiration_seconds_count{job=\"apiserver\"} > 0 and on(job) histogram_quantile(0.01, sum by (job, le) (rate(apiserver_client_certificate_expiration_seconds_bucket{job=\"apiserver\"}[5m]))) < 604800\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeClientCertificateExpiration",
|
|
"annotations": {
|
|
"description": "A client certificate used to authenticate to kubernetes apiserver is expiring in less than 24.0 hours.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeclientcertificateexpiration",
|
|
"summary": "Client certificate is about to expire."
|
|
},
|
|
"expr": "apiserver_client_certificate_expiration_seconds_count{job=\"apiserver\"} > 0 and on(job) histogram_quantile(0.01, sum by (job, le) (rate(apiserver_client_certificate_expiration_seconds_bucket{job=\"apiserver\"}[5m]))) < 86400\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAggregatedAPIErrors",
|
|
"annotations": {
|
|
"description": "Kubernetes aggregated API {{ $labels.name }}/{{ $labels.namespace }} has reported errors. It has appeared unavailable {{ $value | humanize }} times averaged over the past 10m.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeaggregatedapierrors",
|
|
"summary": "Kubernetes aggregated API has reported errors."
|
|
},
|
|
"expr": "sum by(name, namespace, cluster)(increase(aggregator_unavailable_apiservice_total{job=\"apiserver\"}[10m])) > 4\n",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAggregatedAPIDown",
|
|
"annotations": {
|
|
"description": "Kubernetes aggregated API {{ $labels.name }}/{{ $labels.namespace }} has been only {{ $value | humanize }}% available over the last 10m.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeaggregatedapidown",
|
|
"summary": "Kubernetes aggregated API is down."
|
|
},
|
|
"expr": "(1 - max by(name, namespace, cluster)(avg_over_time(aggregator_unavailable_apiservice{job=\"apiserver\"}[10m]))) * 100 < 85\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAPIDown",
|
|
"annotations": {
|
|
"description": "KubeAPI has disappeared from Prometheus target discovery.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapidown",
|
|
"summary": "Target disappeared from Prometheus target discovery."
|
|
},
|
|
"expr": "absent(up{job=\"apiserver\"} == 1)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeAPITerminatedRequests",
|
|
"annotations": {
|
|
"description": "The kubernetes apiserver has terminated {{ $value | humanizePercentage }} of its incoming requests.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeapiterminatedrequests",
|
|
"summary": "The kubernetes apiserver has terminated {{ $value | humanizePercentage }} of its incoming requests."
|
|
},
|
|
"expr": "sum(rate(apiserver_request_terminations_total{job=\"apiserver\"}[10m])) / ( sum(rate(apiserver_request_total{job=\"apiserver\"}[10m])) + sum(rate(apiserver_request_terminations_total{job=\"apiserver\"}[10m])) ) > 0.20\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-system-kubelet",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeNodeNotReady",
|
|
"annotations": {
|
|
"description": "{{ $labels.node }} has been unready for more than 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubenodenotready",
|
|
"summary": "Node is not ready."
|
|
},
|
|
"expr": "kube_node_status_condition{job=\"kube-state-metrics\",condition=\"Ready\",status=\"true\"} == 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeNodeUnreachable",
|
|
"annotations": {
|
|
"description": "{{ $labels.node }} is unreachable and some workloads may be rescheduled.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubenodeunreachable",
|
|
"summary": "Node is unreachable."
|
|
},
|
|
"expr": "(kube_node_spec_taint{job=\"kube-state-metrics\",key=\"node.kubernetes.io/unreachable\",effect=\"NoSchedule\"} unless ignoring(key,value) kube_node_spec_taint{job=\"kube-state-metrics\",key=~\"ToBeDeletedByClusterAutoscaler|cloud.google.com/impending-node-termination|aws-node-termination-handler/spot-itn\"}) == 1\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletTooManyPods",
|
|
"annotations": {
|
|
"description": "Kubelet '{{ $labels.node }}' is running at {{ $value | humanizePercentage }} of its Pod capacity.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubelettoomanypods",
|
|
"summary": "Kubelet is running at capacity."
|
|
},
|
|
"expr": "count by(cluster, node) (\n (kube_pod_status_phase{job=\"kube-state-metrics\",phase=\"Running\"} == 1) * on(instance,pod,namespace,cluster) group_left(node) topk by(instance,pod,namespace,cluster) (1, kube_pod_info{job=\"kube-state-metrics\"})\n)\n/\nmax by(cluster, node) (\n kube_node_status_capacity{job=\"kube-state-metrics\",resource=\"pods\"} != 1\n) > 0.95\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "info"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeNodeReadinessFlapping",
|
|
"annotations": {
|
|
"description": "The readiness status of node {{ $labels.node }} has changed {{ $value }} times in the last 15 minutes.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubenodereadinessflapping",
|
|
"summary": "Node readiness status is flapping."
|
|
},
|
|
"expr": "sum(changes(kube_node_status_condition{job=\"kube-state-metrics\",status=\"true\",condition=\"Ready\"}[15m])) by (cluster, node) > 2\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletPlegDurationHigh",
|
|
"annotations": {
|
|
"description": "The Kubelet Pod Lifecycle Event Generator has a 99th percentile duration of {{ $value }} seconds on node {{ $labels.node }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletplegdurationhigh",
|
|
"summary": "Kubelet Pod Lifecycle Event Generator is taking too long to relist."
|
|
},
|
|
"expr": "node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile{quantile=\"0.99\"} >= 10\n",
|
|
"for": "5m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletPodStartUpLatencyHigh",
|
|
"annotations": {
|
|
"description": "Kubelet Pod startup 99th percentile latency is {{ $value }} seconds on node {{ $labels.node }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletpodstartuplatencyhigh",
|
|
"summary": "Kubelet Pod startup latency is too high."
|
|
},
|
|
"expr": "histogram_quantile(0.99, sum(rate(kubelet_pod_worker_duration_seconds_bucket{job=\"kubelet\", metrics_path=\"/metrics\"}[5m])) by (cluster, instance, le)) * on(cluster, instance) group_left(node) kubelet_node_name{job=\"kubelet\", metrics_path=\"/metrics\"} > 60\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletClientCertificateExpiration",
|
|
"annotations": {
|
|
"description": "Client certificate for Kubelet on node {{ $labels.node }} expires in {{ $value | humanizeDuration }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletclientcertificateexpiration",
|
|
"summary": "Kubelet client certificate is about to expire."
|
|
},
|
|
"expr": "kubelet_certificate_manager_client_ttl_seconds < 604800\n",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletClientCertificateExpiration",
|
|
"annotations": {
|
|
"description": "Client certificate for Kubelet on node {{ $labels.node }} expires in {{ $value | humanizeDuration }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletclientcertificateexpiration",
|
|
"summary": "Kubelet client certificate is about to expire."
|
|
},
|
|
"expr": "kubelet_certificate_manager_client_ttl_seconds < 86400\n",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletServerCertificateExpiration",
|
|
"annotations": {
|
|
"description": "Server certificate for Kubelet on node {{ $labels.node }} expires in {{ $value | humanizeDuration }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletservercertificateexpiration",
|
|
"summary": "Kubelet server certificate is about to expire."
|
|
},
|
|
"expr": "kubelet_certificate_manager_server_ttl_seconds < 604800\n",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletServerCertificateExpiration",
|
|
"annotations": {
|
|
"description": "Server certificate for Kubelet on node {{ $labels.node }} expires in {{ $value | humanizeDuration }}.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletservercertificateexpiration",
|
|
"summary": "Kubelet server certificate is about to expire."
|
|
},
|
|
"expr": "kubelet_certificate_manager_server_ttl_seconds < 86400\n",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletClientCertificateRenewalErrors",
|
|
"annotations": {
|
|
"description": "Kubelet on node {{ $labels.node }} has failed to renew its client certificate ({{ $value | humanize }} errors in the last 5 minutes).",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletclientcertificaterenewalerrors",
|
|
"summary": "Kubelet has failed to renew its client certificate."
|
|
},
|
|
"expr": "increase(kubelet_certificate_manager_client_expiration_renew_errors[5m]) > 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletServerCertificateRenewalErrors",
|
|
"annotations": {
|
|
"description": "Kubelet on node {{ $labels.node }} has failed to renew its server certificate ({{ $value | humanize }} errors in the last 5 minutes).",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletservercertificaterenewalerrors",
|
|
"summary": "Kubelet has failed to renew its server certificate."
|
|
},
|
|
"expr": "increase(kubelet_server_expiration_renew_errors[5m]) > 0\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "warning"
|
|
}
|
|
},
|
|
{
|
|
"alert": "KubeletDown",
|
|
"annotations": {
|
|
"description": "Kubelet has disappeared from Prometheus target discovery.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeletdown",
|
|
"summary": "Target disappeared from Prometheus target discovery."
|
|
},
|
|
"expr": "absent(up{job=\"kubelet\", metrics_path=\"/metrics\"} == 1)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-system-scheduler",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeSchedulerDown",
|
|
"annotations": {
|
|
"description": "KubeScheduler has disappeared from Prometheus target discovery.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubeschedulerdown",
|
|
"summary": "Target disappeared from Prometheus target discovery."
|
|
},
|
|
"expr": "absent(up{job=\"kube-scheduler\"} == 1)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubernetes-system-controller-manager",
|
|
"rules": [
|
|
{
|
|
"alert": "KubeControllerManagerDown",
|
|
"annotations": {
|
|
"description": "KubeControllerManager has disappeared from Prometheus target discovery.",
|
|
"runbook_url": "https://runbooks.prometheus-operator.dev/runbooks/kubernetes/kubecontrollermanagerdown",
|
|
"summary": "Target disappeared from Prometheus target discovery."
|
|
},
|
|
"expr": "absent(up{job=\"kube-controller-manager\"} == 1)\n",
|
|
"for": "15m",
|
|
"labels": {
|
|
"severity": "critical"
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"interval": "3m",
|
|
"name": "kube-apiserver-availability.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "avg_over_time(code_verb:apiserver_request_total:increase1h[30d]) * 24 * 30\n",
|
|
"record": "code_verb:apiserver_request_total:increase30d"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code) (code_verb:apiserver_request_total:increase30d{verb=~\"LIST|GET\"})\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "code:apiserver_request_total:increase30d"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code) (code_verb:apiserver_request_total:increase30d{verb=~\"POST|PUT|PATCH|DELETE\"})\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "code:apiserver_request_total:increase30d"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, verb, scope) (increase(apiserver_request_sli_duration_seconds_count{job=\"apiserver\"}[1h]))\n",
|
|
"record": "cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase1h"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, verb, scope) (avg_over_time(cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase1h[30d]) * 24 * 30)\n",
|
|
"record": "cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, verb, scope, le) (increase(apiserver_request_sli_duration_seconds_bucket[1h]))\n",
|
|
"record": "cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase1h"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, verb, scope, le) (avg_over_time(cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase1h[30d]) * 24 * 30)\n",
|
|
"record": "cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d"
|
|
},
|
|
{
|
|
"expr": "1 - (\n (\n # write too slow\n sum by (cluster) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d{verb=~\"POST|PUT|PATCH|DELETE\"})\n -\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"POST|PUT|PATCH|DELETE\",le=\"1\"})\n ) +\n (\n # read too slow\n sum by (cluster) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d{verb=~\"LIST|GET\"})\n -\n (\n (\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=~\"resource|\",le=\"1\"})\n or\n vector(0)\n )\n +\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=\"namespace\",le=\"5\"})\n +\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=\"cluster\",le=\"30\"})\n )\n ) +\n # errors\n sum by (cluster) (code:apiserver_request_total:increase30d{code=~\"5..\"} or vector(0))\n)\n/\nsum by (cluster) (code:apiserver_request_total:increase30d)\n",
|
|
"labels": {
|
|
"verb": "all"
|
|
},
|
|
"record": "apiserver_request:availability30d"
|
|
},
|
|
{
|
|
"expr": "1 - (\n sum by (cluster) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d{verb=~\"LIST|GET\"})\n -\n (\n # too slow\n (\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=~\"resource|\",le=\"1\"})\n or\n vector(0)\n )\n +\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=\"namespace\",le=\"5\"})\n +\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"LIST|GET\",scope=\"cluster\",le=\"30\"})\n )\n +\n # errors\n sum by (cluster) (code:apiserver_request_total:increase30d{verb=\"read\",code=~\"5..\"} or vector(0))\n)\n/\nsum by (cluster) (code:apiserver_request_total:increase30d{verb=\"read\"})\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:availability30d"
|
|
},
|
|
{
|
|
"expr": "1 - (\n (\n # too slow\n sum by (cluster) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase30d{verb=~\"POST|PUT|PATCH|DELETE\"})\n -\n sum by (cluster) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase30d{verb=~\"POST|PUT|PATCH|DELETE\",le=\"1\"})\n )\n +\n # errors\n sum by (cluster) (code:apiserver_request_total:increase30d{verb=\"write\",code=~\"5..\"} or vector(0))\n)\n/\nsum by (cluster) (code:apiserver_request_total:increase30d{verb=\"write\"})\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:availability30d"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster,code,resource) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[5m]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "code_resource:apiserver_request_total:rate5m"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster,code,resource) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[5m]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "code_resource:apiserver_request_total:rate5m"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code, verb) (increase(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET|POST|PUT|PATCH|DELETE\",code=~\"2..\"}[1h]))\n",
|
|
"record": "code_verb:apiserver_request_total:increase1h"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code, verb) (increase(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET|POST|PUT|PATCH|DELETE\",code=~\"3..\"}[1h]))\n",
|
|
"record": "code_verb:apiserver_request_total:increase1h"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code, verb) (increase(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET|POST|PUT|PATCH|DELETE\",code=~\"4..\"}[1h]))\n",
|
|
"record": "code_verb:apiserver_request_total:increase1h"
|
|
},
|
|
{
|
|
"expr": "sum by (cluster, code, verb) (increase(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET|POST|PUT|PATCH|DELETE\",code=~\"5..\"}[1h]))\n",
|
|
"record": "code_verb:apiserver_request_total:increase1h"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kube-apiserver-burnrate.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[1d]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[1d]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[1d]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[1d]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[1d]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[1d]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate1d"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[1h]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[1h]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[1h]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[1h]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[1h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[1h]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate1h"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[2h]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[2h]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[2h]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[2h]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[2h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[2h]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate2h"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[30m]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[30m]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[30m]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[30m]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[30m]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[30m]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate30m"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[3d]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[3d]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[3d]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[3d]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[3d]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[3d]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate3d"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[5m]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[5m]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[5m]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[5m]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[5m]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[5m]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate5m"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[6h]))\n -\n (\n (\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=~\"resource|\",le=\"1\"}[6h]))\n or\n vector(0)\n )\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"namespace\",le=\"5\"}[6h]))\n +\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\",scope=\"cluster\",le=\"30\"}[6h]))\n )\n )\n +\n # errors\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\",code=~\"5..\"}[6h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"LIST|GET\"}[6h]))\n",
|
|
"labels": {
|
|
"verb": "read"
|
|
},
|
|
"record": "apiserver_request:burnrate6h"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[1d]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[1d]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[1d]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[1d]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate1d"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[1h]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[1h]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[1h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[1h]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate1h"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[2h]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[2h]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[2h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[2h]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate2h"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[30m]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[30m]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[30m]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[30m]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate30m"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[3d]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[3d]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[3d]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[3d]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate3d"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[5m]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[5m]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[5m]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[5m]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate5m"
|
|
},
|
|
{
|
|
"expr": "(\n (\n # too slow\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_count{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[6h]))\n -\n sum by (cluster) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\",le=\"1\"}[6h]))\n )\n +\n sum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",code=~\"5..\"}[6h]))\n)\n/\nsum by (cluster) (rate(apiserver_request_total{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\"}[6h]))\n",
|
|
"labels": {
|
|
"verb": "write"
|
|
},
|
|
"record": "apiserver_request:burnrate6h"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kube-apiserver-histogram.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum by (cluster, le, resource) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"LIST|GET\",subresource!~\"proxy|attach|log|exec|portforward\"}[5m]))) > 0\n",
|
|
"labels": {
|
|
"quantile": "0.99",
|
|
"verb": "read"
|
|
},
|
|
"record": "cluster_quantile:apiserver_request_sli_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum by (cluster, le, resource) (rate(apiserver_request_sli_duration_seconds_bucket{job=\"apiserver\",verb=~\"POST|PUT|PATCH|DELETE\",subresource!~\"proxy|attach|log|exec|portforward\"}[5m]))) > 0\n",
|
|
"labels": {
|
|
"quantile": "0.99",
|
|
"verb": "write"
|
|
},
|
|
"record": "cluster_quantile:apiserver_request_sli_duration_seconds:histogram_quantile"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_cpu_usage_seconds_total",
|
|
"rules": [
|
|
{
|
|
"expr": "sum by (cluster, namespace, pod, container) (\n irate(container_cpu_usage_seconds_total{job=\"kubelet\", metrics_path=\"/metrics/cadvisor\", image!=\"\"}[5m])\n) * on (cluster, namespace, pod) group_left(node) topk by (cluster, namespace, pod) (\n 1, max by(cluster, namespace, pod, node) (kube_pod_info{node!=\"\"})\n)\n",
|
|
"record": "node_namespace_pod_container:container_cpu_usage_seconds_total:sum_irate"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_memory_working_set_bytes",
|
|
"rules": [
|
|
{
|
|
"expr": "container_memory_working_set_bytes{job=\"kubelet\", metrics_path=\"/metrics/cadvisor\", image!=\"\"}\n* on (cluster, namespace, pod) group_left(node) topk by(cluster, namespace, pod) (1,\n max by(cluster, namespace, pod, node) (kube_pod_info{node!=\"\"})\n)\n",
|
|
"record": "node_namespace_pod_container:container_memory_working_set_bytes"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_memory_rss",
|
|
"rules": [
|
|
{
|
|
"expr": "container_memory_rss{job=\"kubelet\", metrics_path=\"/metrics/cadvisor\", image!=\"\"}\n* on (cluster, namespace, pod) group_left(node) topk by(cluster, namespace, pod) (1,\n max by(cluster, namespace, pod, node) (kube_pod_info{node!=\"\"})\n)\n",
|
|
"record": "node_namespace_pod_container:container_memory_rss"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_memory_cache",
|
|
"rules": [
|
|
{
|
|
"expr": "container_memory_cache{job=\"kubelet\", metrics_path=\"/metrics/cadvisor\", image!=\"\"}\n* on (cluster, namespace, pod) group_left(node) topk by(cluster, namespace, pod) (1,\n max by(cluster, namespace, pod, node) (kube_pod_info{node!=\"\"})\n)\n",
|
|
"record": "node_namespace_pod_container:container_memory_cache"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_memory_swap",
|
|
"rules": [
|
|
{
|
|
"expr": "container_memory_swap{job=\"kubelet\", metrics_path=\"/metrics/cadvisor\", image!=\"\"}\n* on (cluster, namespace, pod) group_left(node) topk by(cluster, namespace, pod) (1,\n max by(cluster, namespace, pod, node) (kube_pod_info{node!=\"\"})\n)\n",
|
|
"record": "node_namespace_pod_container:container_memory_swap"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.container_resource",
|
|
"rules": [
|
|
{
|
|
"expr": "kube_pod_container_resource_requests{resource=\"memory\",job=\"kube-state-metrics\"} * on (namespace, pod, cluster)\ngroup_left() max by (namespace, pod, cluster) (\n (kube_pod_status_phase{phase=~\"Pending|Running\"} == 1)\n)\n",
|
|
"record": "cluster:namespace:pod_memory:active:kube_pod_container_resource_requests"
|
|
},
|
|
{
|
|
"expr": "sum by (namespace, cluster) (\n sum by (namespace, pod, cluster) (\n max by (namespace, pod, container, cluster) (\n kube_pod_container_resource_requests{resource=\"memory\",job=\"kube-state-metrics\"}\n ) * on(namespace, pod, cluster) group_left() max by (namespace, pod, cluster) (\n kube_pod_status_phase{phase=~\"Pending|Running\"} == 1\n )\n )\n)\n",
|
|
"record": "namespace_memory:kube_pod_container_resource_requests:sum"
|
|
},
|
|
{
|
|
"expr": "kube_pod_container_resource_requests{resource=\"cpu\",job=\"kube-state-metrics\"} * on (namespace, pod, cluster)\ngroup_left() max by (namespace, pod, cluster) (\n (kube_pod_status_phase{phase=~\"Pending|Running\"} == 1)\n)\n",
|
|
"record": "cluster:namespace:pod_cpu:active:kube_pod_container_resource_requests"
|
|
},
|
|
{
|
|
"expr": "sum by (namespace, cluster) (\n sum by (namespace, pod, cluster) (\n max by (namespace, pod, container, cluster) (\n kube_pod_container_resource_requests{resource=\"cpu\",job=\"kube-state-metrics\"}\n ) * on(namespace, pod, cluster) group_left() max by (namespace, pod, cluster) (\n kube_pod_status_phase{phase=~\"Pending|Running\"} == 1\n )\n )\n)\n",
|
|
"record": "namespace_cpu:kube_pod_container_resource_requests:sum"
|
|
},
|
|
{
|
|
"expr": "kube_pod_container_resource_limits{resource=\"memory\",job=\"kube-state-metrics\"} * on (namespace, pod, cluster)\ngroup_left() max by (namespace, pod, cluster) (\n (kube_pod_status_phase{phase=~\"Pending|Running\"} == 1)\n)\n",
|
|
"record": "cluster:namespace:pod_memory:active:kube_pod_container_resource_limits"
|
|
},
|
|
{
|
|
"expr": "sum by (namespace, cluster) (\n sum by (namespace, pod, cluster) (\n max by (namespace, pod, container, cluster) (\n kube_pod_container_resource_limits{resource=\"memory\",job=\"kube-state-metrics\"}\n ) * on(namespace, pod, cluster) group_left() max by (namespace, pod, cluster) (\n kube_pod_status_phase{phase=~\"Pending|Running\"} == 1\n )\n )\n)\n",
|
|
"record": "namespace_memory:kube_pod_container_resource_limits:sum"
|
|
},
|
|
{
|
|
"expr": "kube_pod_container_resource_limits{resource=\"cpu\",job=\"kube-state-metrics\"} * on (namespace, pod, cluster)\ngroup_left() max by (namespace, pod, cluster) (\n (kube_pod_status_phase{phase=~\"Pending|Running\"} == 1)\n )\n",
|
|
"record": "cluster:namespace:pod_cpu:active:kube_pod_container_resource_limits"
|
|
},
|
|
{
|
|
"expr": "sum by (namespace, cluster) (\n sum by (namespace, pod, cluster) (\n max by (namespace, pod, container, cluster) (\n kube_pod_container_resource_limits{resource=\"cpu\",job=\"kube-state-metrics\"}\n ) * on(namespace, pod, cluster) group_left() max by (namespace, pod, cluster) (\n kube_pod_status_phase{phase=~\"Pending|Running\"} == 1\n )\n )\n)\n",
|
|
"record": "namespace_cpu:kube_pod_container_resource_limits:sum"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "k8s.rules.pod_owner",
|
|
"rules": [
|
|
{
|
|
"expr": "max by (cluster, namespace, workload, pod) (\n label_replace(\n label_replace(\n kube_pod_owner{job=\"kube-state-metrics\", owner_kind=\"ReplicaSet\"},\n \"replicaset\", \"$1\", \"owner_name\", \"(.*)\"\n ) * on(replicaset, namespace) group_left(owner_name) topk by(replicaset, namespace) (\n 1, max by (replicaset, namespace, owner_name) (\n kube_replicaset_owner{job=\"kube-state-metrics\"}\n )\n ),\n \"workload\", \"$1\", \"owner_name\", \"(.*)\"\n )\n)\n",
|
|
"labels": {
|
|
"workload_type": "deployment"
|
|
},
|
|
"record": "namespace_workload_pod:kube_pod_owner:relabel"
|
|
},
|
|
{
|
|
"expr": "max by (cluster, namespace, workload, pod) (\n label_replace(\n kube_pod_owner{job=\"kube-state-metrics\", owner_kind=\"DaemonSet\"},\n \"workload\", \"$1\", \"owner_name\", \"(.*)\"\n )\n)\n",
|
|
"labels": {
|
|
"workload_type": "daemonset"
|
|
},
|
|
"record": "namespace_workload_pod:kube_pod_owner:relabel"
|
|
},
|
|
{
|
|
"expr": "max by (cluster, namespace, workload, pod) (\n label_replace(\n kube_pod_owner{job=\"kube-state-metrics\", owner_kind=\"StatefulSet\"},\n \"workload\", \"$1\", \"owner_name\", \"(.*)\"\n )\n)\n",
|
|
"labels": {
|
|
"workload_type": "statefulset"
|
|
},
|
|
"record": "namespace_workload_pod:kube_pod_owner:relabel"
|
|
},
|
|
{
|
|
"expr": "max by (cluster, namespace, workload, pod) (\n label_replace(\n kube_pod_owner{job=\"kube-state-metrics\", owner_kind=\"Job\"},\n \"workload\", \"$1\", \"owner_name\", \"(.*)\"\n )\n)\n",
|
|
"labels": {
|
|
"workload_type": "job"
|
|
},
|
|
"record": "namespace_workload_pod:kube_pod_owner:relabel"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kube-scheduler.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum(rate(scheduler_e2e_scheduling_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.99"
|
|
},
|
|
"record": "cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum(rate(scheduler_scheduling_algorithm_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.99"
|
|
},
|
|
"record": "cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum(rate(scheduler_binding_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.99"
|
|
},
|
|
"record": "cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.9, sum(rate(scheduler_e2e_scheduling_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.9"
|
|
},
|
|
"record": "cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.9, sum(rate(scheduler_scheduling_algorithm_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.9"
|
|
},
|
|
"record": "cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.9, sum(rate(scheduler_binding_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.9"
|
|
},
|
|
"record": "cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.5, sum(rate(scheduler_e2e_scheduling_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.5"
|
|
},
|
|
"record": "cluster_quantile:scheduler_e2e_scheduling_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.5, sum(rate(scheduler_scheduling_algorithm_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.5"
|
|
},
|
|
"record": "cluster_quantile:scheduler_scheduling_algorithm_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.5, sum(rate(scheduler_binding_duration_seconds_bucket{job=\"kube-scheduler\"}[5m])) without(instance, pod))\n",
|
|
"labels": {
|
|
"quantile": "0.5"
|
|
},
|
|
"record": "cluster_quantile:scheduler_binding_duration_seconds:histogram_quantile"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "node.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "topk by(cluster, namespace, pod) (1,\n max by (cluster, node, namespace, pod) (\n label_replace(kube_pod_info{job=\"kube-state-metrics\",node!=\"\"}, \"pod\", \"$1\", \"pod\", \"(.*)\")\n))\n",
|
|
"record": "node_namespace_pod:kube_pod_info:"
|
|
},
|
|
{
|
|
"expr": "count by (cluster, node) (\n node_cpu_seconds_total{mode=\"idle\",job=\"node-exporter\"}\n * on (cluster, namespace, pod) group_left(node)\n topk by(cluster, namespace, pod) (1, node_namespace_pod:kube_pod_info:)\n)\n",
|
|
"record": "node:node_num_cpu:sum"
|
|
},
|
|
{
|
|
"expr": "sum(\n node_memory_MemAvailable_bytes{job=\"node-exporter\"} or\n (\n node_memory_Buffers_bytes{job=\"node-exporter\"} +\n node_memory_Cached_bytes{job=\"node-exporter\"} +\n node_memory_MemFree_bytes{job=\"node-exporter\"} +\n node_memory_Slab_bytes{job=\"node-exporter\"}\n )\n) by (cluster)\n",
|
|
"record": ":node_memory_MemAvailable_bytes:sum"
|
|
},
|
|
{
|
|
"expr": "avg by (cluster, node) (\n sum without (mode) (\n rate(node_cpu_seconds_total{mode!=\"idle\",mode!=\"iowait\",mode!=\"steal\",job=\"node-exporter\"}[5m])\n )\n)\n",
|
|
"record": "node:node_cpu_utilization:ratio_rate5m"
|
|
},
|
|
{
|
|
"expr": "avg by (cluster) (\n node:node_cpu_utilization:ratio_rate5m\n)\n",
|
|
"record": "cluster:node_cpu:ratio_rate5m"
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"name": "kubelet.rules",
|
|
"rules": [
|
|
{
|
|
"expr": "histogram_quantile(0.99, sum(rate(kubelet_pleg_relist_duration_seconds_bucket{job=\"kubelet\", metrics_path=\"/metrics\"}[5m])) by (cluster, instance, le) * on(cluster, instance) group_left(node) kubelet_node_name{job=\"kubelet\", metrics_path=\"/metrics\"})\n",
|
|
"labels": {
|
|
"quantile": "0.99"
|
|
},
|
|
"record": "node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.9, sum(rate(kubelet_pleg_relist_duration_seconds_bucket{job=\"kubelet\", metrics_path=\"/metrics\"}[5m])) by (cluster, instance, le) * on(cluster, instance) group_left(node) kubelet_node_name{job=\"kubelet\", metrics_path=\"/metrics\"})\n",
|
|
"labels": {
|
|
"quantile": "0.9"
|
|
},
|
|
"record": "node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile"
|
|
},
|
|
{
|
|
"expr": "histogram_quantile(0.5, sum(rate(kubelet_pleg_relist_duration_seconds_bucket{job=\"kubelet\", metrics_path=\"/metrics\"}[5m])) by (cluster, instance, le) * on(cluster, instance) group_left(node) kubelet_node_name{job=\"kubelet\", metrics_path=\"/metrics\"})\n",
|
|
"labels": {
|
|
"quantile": "0.5"
|
|
},
|
|
"record": "node_quantile:kubelet_pleg_relist_duration_seconds:histogram_quantile"
|
|
}
|
|
]
|
|
}
|
|
]
|
|
}
|
|
}
|