Default Prometheus metrics configuration in Azure Monitor

This article lists the default targets, dashboards, and recording rules when you configure Prometheus metrics to be scraped from an Azure Kubernetes Service (AKS) cluster for any AKS cluster.

Minimal ingestion profile

Minimal ingestion profile is a setting that helps reduce ingestion volume of metrics, as only metrics used by default dashboards, default recording rules & default alerts are collected. For addon based collection, Minimal ingestion profile setting is enabled by default. You can modify collection to enable collecting more metrics, as specified below.

Scrape frequency

The default scrape frequency for all default targets and scrapes is 30 seconds.

Targets scraped by default

Following targets are enabled/ON by default - meaning you don't have to provide any scrape job configuration for scraping these targets, as metrics addon will scrape these targets automatically by default.

• cadvisor (job=cadvisor)
• nodeexporter (job=node)
• kubelet (job=kubelet)
• kube-state-metrics (job=kube-state-metrics)
• networkobservabilityRetina (job=networkobservabilityRetina)

The following targets are enabled/ON when you enable control plane metrics (preview) feature. You can use control plane metrics to maximize overall observability and maintain operational excellence for your AKS cluster. For more information, see Control plane metrics (preview).

• controlplane-apiserver (job=controlplane-apiserver)
• controlplane-etcd (job=controlplane-etcd)

The following targets are enabled/ON when you enable Container Network Observability which is a feature of the Advanced Container Networking Services suite and is compatible with all Linux workloads seamlessly integrating with Hubble for both Cilium or non-Cilium based data plane. This allows flexibility for your container networking needs. For more information, see Advanced Container Networking Services.

• networkobservabilityHubble (job=networkobservabilityHubble)
• networkobservabilityCilium (job=networkobservabilityCilium)

Metrics collected from default targets

The following metrics are collected by default from each default target. All other metrics are dropped through relabeling rules.

cadvisor (job=cadvisor)

• container_spec_cpu_period
• container_spec_cpu_quota
• container_cpu_usage_seconds_total
• container_memory_rss
• container_network_receive_bytes_total
• container_network_transmit_bytes_total
• container_network_receive_packets_total
• container_network_transmit_packets_total
• container_network_receive_packets_dropped_total
• container_network_transmit_packets_dropped_total
• container_fs_reads_total
• container_fs_writes_total
• container_fs_reads_bytes_total
• container_fs_writes_bytes_total
• container_memory_working_set_bytes
• container_memory_cache
• container_memory_swap
• container_cpu_cfs_throttled_periods_total
• container_cpu_cfs_periods_total
• container_memory_usage_bytes
• kubernetes_build_info"

kubelet (job=kubelet)

• kubelet_volume_stats_used_bytes
• kubelet_node_name
• kubelet_running_pods
• kubelet_running_pod_count
• kubelet_running_containers
• kubelet_running_container_count
• volume_manager_total_volumes
• kubelet_node_config_error
• kubelet_runtime_operations_total
• kubelet_runtime_operations_errors_total
• kubelet_runtime_operations_duration_seconds kubelet_runtime_operations_duration_seconds_bucket kubelet_runtime_operations_duration_seconds_sum kubelet_runtime_operations_duration_seconds_count
• kubelet_pod_start_duration_seconds kubelet_pod_start_duration_seconds_bucket kubelet_pod_start_duration_seconds_sum kubelet_pod_start_duration_seconds_count
• kubelet_pod_worker_duration_seconds kubelet_pod_worker_duration_seconds_bucket kubelet_pod_worker_duration_seconds_sum kubelet_pod_worker_duration_seconds_count
• storage_operation_duration_seconds storage_operation_duration_seconds_bucket storage_operation_duration_seconds_sum storage_operation_duration_seconds_count
• storage_operation_errors_total
• kubelet_cgroup_manager_duration_seconds kubelet_cgroup_manager_duration_seconds_bucket kubelet_cgroup_manager_duration_seconds_sum kubelet_cgroup_manager_duration_seconds_count
• kubelet_pleg_relist_duration_seconds kubelet_pleg_relist_duration_seconds_bucket kubelet_pleg_relist_duration_sum kubelet_pleg_relist_duration_seconds_count
• kubelet_pleg_relist_interval_seconds kubelet_pleg_relist_interval_seconds_bucket kubelet_pleg_relist_interval_seconds_sum kubelet_pleg_relist_interval_seconds_count
• rest_client_requests_total
• rest_client_request_duration_seconds rest_client_request_duration_seconds_bucket rest_client_request_duration_seconds_sum rest_client_request_duration_seconds_count
• process_resident_memory_bytes
• process_cpu_seconds_total
• go_goroutines
• kubelet_volume_stats_capacity_bytes
• kubelet_volume_stats_available_bytes
• kubelet_volume_stats_inodes_used
• kubelet_volume_stats_inodes
• kubernetes_build_info"

nodexporter (job=node)

• node_cpu_seconds_total
• node_memory_MemAvailable_bytes
• node_memory_Buffers_bytes
• node_memory_Cached_bytes
• node_memory_MemFree_bytes
• node_memory_Slab_bytes
• node_memory_MemTotal_bytes
• node_netstat_Tcp_RetransSegs
• node_netstat_Tcp_OutSegs
• node_netstat_TcpExt_TCPSynRetrans
• node_load1``node_load5
• node_load15
• node_disk_read_bytes_total
• node_disk_written_bytes_total
• node_disk_io_time_seconds_total
• node_filesystem_size_bytes
• node_filesystem_avail_bytes
• node_filesystem_readonly
• node_network_receive_bytes_total
• node_network_transmit_bytes_total
• node_vmstat_pgmajfault
• node_network_receive_drop_total
• node_network_transmit_drop_total
• node_disk_io_time_weighted_seconds_total
• node_exporter_build_info
• node_time_seconds
• node_uname_info"

kube-state-metrics (job=kube-state-metrics)

• kube_job_status_succeeded
• kube_job_spec_completions
• kube_daemonset_status_desired_number_scheduled
• kube_daemonset_status_number_ready
• kube_deployment_status_replicas_ready
• kube_pod_container_status_last_terminated_reason
• kube_pod_container_status_waiting_reason
• kube_pod_container_status_restarts_total
• kube_node_status_allocatable
• kube_pod_owner
• kube_pod_container_resource_requests
• kube_pod_status_phase
• kube_pod_container_resource_limits
• kube_replicaset_owner
• kube_resourcequota
• kube_namespace_status_phase
• kube_node_status_capacity
• kube_node_info
• kube_pod_info
• kube_deployment_spec_replicas
• kube_deployment_status_replicas_available
• kube_deployment_status_replicas_updated
• kube_statefulset_status_replicas_ready
• kube_statefulset_status_replicas
• kube_statefulset_status_replicas_updated
• kube_job_status_start_time
• kube_job_status_active
• kube_job_failed
• kube_horizontalpodautoscaler_status_desired_replicas
• kube_horizontalpodautoscaler_status_current_replicas
• kube_horizontalpodautoscaler_spec_min_replicas
• kube_horizontalpodautoscaler_spec_max_replicas
• kubernetes_build_info
• kube_node_status_condition
• kube_node_spec_taint
• kube_pod_container_info
• kube_resource_labels (ex - kube_pod_labels, kube_deployment_labels)
• kube_resource_annotations (ex - kube_pod_annotations, kube_deployment_annotations)

controlplane-apiserver (job=controlplane-apiserver)

• apiserver_request_total
• apiserver_cache_list_fetched_objects_total
• apiserver_cache_list_returned_objects_total
• apiserver_flowcontrol_demand_seats_average
• apiserver_flowcontrol_current_limit_seats
• apiserver_request_sli_duration_seconds_bucket
• apiserver_request_sli_duration_seconds_count
• apiserver_request_sli_duration_seconds_sum
• process_start_time_seconds
• apiserver_request_duration_seconds_bucket
• apiserver_request_duration_seconds_count
• apiserver_request_duration_seconds_sum
• apiserver_storage_list_fetched_objects_total
• apiserver_storage_list_returned_objects_total
• apiserver_current_inflight_requests

controlplane-etcd (job=controlplane-etcd)

• etcd_server_has_leader
• rest_client_requests_total
• etcd_mvcc_db_total_size_in_bytes
• etcd_mvcc_db_total_size_in_use_in_bytes
• etcd_server_slow_read_indexes_total
• etcd_server_slow_apply_total
• etcd_network_client_grpc_sent_bytes_total
• etcd_server_heartbeat_send_failures_total

networkobservabilityHubble (job=networkobservabilityHubble), and networkobservabilityCilium (job=networkobservabilityCilium)
For list of metrics collected by these targets, see Container Network Observability metrics

Default targets scraped for Windows

Following Windows targets are configured to scrape, but scraping isn't enabled (disabled/OFF) by default - meaning you don't have to provide any scrape job configuration for scraping these targets but they are disabled/OFF by default and you need to turn ON/enable scraping for these targets using ama-metrics-settings-configmap under default-scrape-settings-enabled section.

Two default jobs can be run for Windows that scrape metrics required for the dashboards specific to Windows.

• windows-exporter (job=windows-exporter)
• kube-proxy-windows (job=kube-proxy-windows)

Note

This requires applying or updating the ama-metrics-settings-configmap configmap and installing windows-exporter on all Windows nodes. For more information, see the enablement document.

Metrics scraped for Windows

The following metrics are collected when windows-exporter and kube-proxy-windows are enabled.

windows-exporter (job=windows-exporter)

• windows_system_system_up_time
• windows_cpu_time_total
• windows_memory_available_bytes
• windows_os_visible_memory_bytes
• windows_memory_cache_bytes
• windows_memory_modified_page_list_bytes
• windows_memory_standby_cache_core_bytes
• windows_memory_standby_cache_normal_priority_bytes
• windows_memory_standby_cache_reserve_bytes
• windows_memory_swap_page_operations_total
• windows_logical_disk_read_seconds_total
• windows_logical_disk_write_seconds_total
• windows_logical_disk_size_bytes
• windows_logical_disk_free_bytes
• windows_net_bytes_total
• windows_net_packets_received_discarded_total
• windows_net_packets_outbound_discarded_total
• windows_container_available
• windows_container_cpu_usage_seconds_total
• windows_container_memory_usage_commit_bytes
• windows_container_memory_usage_private_working_set_bytes
• windows_container_network_receive_bytes_total
• windows_container_network_transmit_bytes_total

kube-proxy-windows (job=kube-proxy-windows)

• kubeproxy_sync_proxy_rules_duration_seconds
• kubeproxy_sync_proxy_rules_duration_seconds_bucket
• kubeproxy_sync_proxy_rules_duration_seconds_sum
• kubeproxy_sync_proxy_rules_duration_seconds_count
• rest_client_requests_total
• rest_client_request_duration_seconds
• rest_client_request_duration_seconds_bucket
• rest_client_request_duration_seconds_sum
• rest_client_request_duration_seconds_count
• process_resident_memory_bytes
• process_cpu_seconds_total
• go_goroutines

Dashboards

The following default dashboards are automatically provisioned and configured by Azure Monitor managed service for Prometheus when you link your Azure Monitor workspace to an Azure Managed Grafana instance. Source code for these dashboards can be found in this GitHub repository. The below dashboards will be provisioned in the specified Azure Grafana instance under Managed Prometheus folder in Grafana. These are the standard open source community dashboards for monitoring Kubernetes clusters with Prometheus and Grafana.

• Kubernetes / Compute Resources / Cluster
• Kubernetes / Compute Resources / Namespace (Pods)
• Kubernetes / Compute Resources / Node (Pods)
• Kubernetes / Compute Resources / Pod
• Kubernetes / Compute Resources / Namespace (Workloads)
• Kubernetes / Compute Resources / Workload
• Kubernetes / Kubelet
• Node Exporter / USE Method / Node
• Node Exporter / Nodes
• Kubernetes / Compute Resources / Cluster (Windows)
• Kubernetes / Compute Resources / Namespace (Windows)
• Kubernetes / Compute Resources / Pod (Windows)
• Kubernetes / USE Method / Cluster (Windows)
• Kubernetes / USE Method / Node (Windows)

Recording rules

The following default recording rules are automatically configured by Azure Monitor managed service for Prometheus when you configure Prometheus metrics to be scraped from an Azure Kubernetes Service (AKS) cluster. Source code for these recording rules can be found in this GitHub repository. These are the standard open source recording rules used in the dashboards above.

• cluster:node_cpu:ratio_rate5m
• namespace_cpu:kube_pod_container_resource_requests:sum
• namespace_cpu:kube_pod_container_resource_limits:sum
• :node_memory_MemAvailable_bytes:sum
• namespace_memory:kube_pod_container_resource_requests:sum
• namespace_memory:kube_pod_container_resource_limits:sum
• namespace_workload_pod:kube_pod_owner:relabel
• node_namespace_pod_container:container_cpu_usage_seconds_total:sum_irate
• cluster:namespace:pod_cpu:active:kube_pod_container_resource_requests
• cluster:namespace:pod_cpu:active:kube_pod_container_resource_limits
• cluster:namespace:pod_memory:active:kube_pod_container_resource_requests
• cluster:namespace:pod_memory:active:kube_pod_container_resource_limits
• node_namespace_pod_container:container_memory_working_set_bytes
• node_namespace_pod_container:container_memory_rss
• node_namespace_pod_container:container_memory_cache
• node_namespace_pod_container:container_memory_swap
• instance:node_cpu_utilisation:rate5m
• instance:node_load1_per_cpu:ratio
• instance:node_memory_utilisation:ratio
• instance:node_vmstat_pgmajfault:rate5m
• instance:node_network_receive_bytes_excluding_lo:rate5m
• instance:node_network_transmit_bytes_excluding_lo:rate5m
• instance:node_network_receive_drop_excluding_lo:rate5m
• instance:node_network_transmit_drop_excluding_lo:rate5m
• instance_device:node_disk_io_time_seconds:rate5m
• instance_device:node_disk_io_time_weighted_seconds:rate5m
• instance:node_num_cpu:sum
• node:windows_node:sum
• node:windows_node_num_cpu:sum
• :windows_node_cpu_utilisation:avg5m
• node:windows_node_cpu_utilisation:avg5m
• :windows_node_memory_utilisation:
• :windows_node_memory_MemFreeCached_bytes:sum
• node:windows_node_memory_totalCached_bytes:sum
• :windows_node_memory_MemTotal_bytes:sum
• node:windows_node_memory_bytes_available:sum
• node:windows_node_memory_bytes_total:sum
• node:windows_node_memory_utilisation:ratio
• node:windows_node_memory_utilisation:
• node:windows_node_memory_swap_io_pages:irate
• :windows_node_disk_utilisation:avg_irate
• node:windows_node_disk_utilisation:avg_irate
• node:windows_node_filesystem_usage:
• node:windows_node_filesystem_avail:
• :windows_node_net_utilisation:sum_irate
• node:windows_node_net_utilisation:sum_irate
• :windows_node_net_saturation:sum_irate
• node:windows_node_net_saturation:sum_irate
• windows_pod_container_available
• windows_container_total_runtime
• windows_container_memory_usage
• windows_container_private_working_set_usage
• windows_container_network_received_bytes_total
• windows_container_network_transmitted_bytes_total
• kube_pod_windows_container_resource_memory_request
• kube_pod_windows_container_resource_memory_limit
• kube_pod_windows_container_resource_cpu_cores_request
• kube_pod_windows_container_resource_cpu_cores_limit
• namespace_pod_container:windows_container_cpu_usage_seconds_total:sum_rate

Prometheus visualization recording rules

When you use Prometheus based Container Insights, more recording rules will be deployed to support the Prometheus visualizations.

• ux:cluster_pod_phase_count:sum
• ux:node_cpu_usage:sum_irate
• ux:node_memory_usage:sum
• ux:controller_pod_phase_count:sum
• ux:controller_container_count:sum
• ux:controller_workingset_memory:sum
• ux:controller_cpu_usage:sum_irate
• ux:controller_rss_memory:sum
• ux:controller_resource_limit:sum
• ux:controller_container_restarts:max
• ux:pod_container_count:sum
• ux:pod_cpu_usage:sum_irate
• ux:pod_workingset_memory:sum
• ux:pod_rss_memory:sum
• ux:pod_resource_limit:sum
• ux:pod_container_restarts:max
• ux:node_network_receive_drop_total:sum_irate
• ux:node_network_transmit_drop_total:sum_irate

The following recording rules are required for Windows support. These rules are deployed along with the above rules, however, they aren't enabled by default. Follow the instructions for enabling and disabling rule groups in your Azure Monitor workspace.

• ux:node_cpu_usage_windows:sum_irate
• ux:node_memory_usage_windows:sum
• ux:controller_cpu_usage_windows:sum_irate
• ux:controller_workingset_memory_windows:sum
• ux:pod_cpu_usage_windows:sum_irate
• ux:pod_workingset_memory_windows:sum

Next steps

Customize scraping of Prometheus metrics.