Skip to main content
Version: v1.7

Karmada Metrics Reference

Metrics

This section details the metrics that different Karmada components export. You can query the metrics endpoint for these components using an HTTP scrape, and fetch the current metrics data in Prometheus format.

NameTypeHelpLabelsSource Components
schedule_attempts_totalCounterNumber of attempts to schedule resourceBinding.result
schedule_type
karmada-scheduler
e2e_scheduling_duration_secondsHistogramE2E scheduling latency in seconds.result
schedule_type
karmada-scheduler
scheduling_algorithm_duration_secondsHistogramScheduling algorithm latency in seconds(excluding scale scheduler).schedule_stepkarmada-scheduler
queue_incoming_bindings_totalCounterNumber of bindings added to scheduling queues by event type.eventkarmada-scheduler
framework_extension_point_duration_secondsHistogramLatency for running all plugins of a specific extension point.extension_point
result
karmada-scheduler
plugin_execution_duration_secondsHistogramDuration for running a plugin at a specific extension point.plugin
extension_point
result
karmada-scheduler
estimating_request_totalCounterNumber of scheduler estimator requests.result
type
karmada_scheduler_estimator
estimating_algorithm_duration_secondsHistogramEstimating algorithm latency in seconds for each step.result
type
step
karmada_scheduler_estimator
cluster_ready_stateGaugeState of the cluster(1 if ready, 0 otherwise).cluster_namekarmada-controller-manager
karmada-agent
cluster_node_numberGaugeNumber of nodes in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_ready_node_numberGaugeNumber of ready nodes in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_memory_allocatable_bytesGaugeAllocatable cluster memory resource in bytes.cluster_namekarmada-controller-manager
karmada-agent
cluster_cpu_allocatable_numberGaugeNumber of allocatable CPU in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_pod_allocatable_numberGaugeNumber of allocatable pods in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_memory_allocated_bytesGaugeAllocated cluster memory resource in bytes.cluster_namekarmada-controller-manager
karmada-agent
cluster_cpu_allocated_numberGaugeNumber of allocated CPU in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_pod_allocated_numberGaugeNumber of allocated pods in the cluster.cluster_namekarmada-controller-manager
karmada-agent
cluster_sync_status_duration_secondsHistogramDuration in seconds for syncing the status of the cluster once.cluster_namekarmada-controller-manager
karmada-agent
resource_match_policy_duration_secondsHistogramDuration in seconds to find a matched propagation policy for the resource template./karmada-controller-manager
resource_apply_policy_duration_secondsHistogramDuration in seconds to apply a propagation policy for the resource template. By the result, 'error' means a resource template failed to apply the policy. Otherwise 'success'.resultkarmada-controller-manager
policy_apply_attempts_totalCounterNumber of attempts to be applied for a propagation policy. By the result, 'error' means a resource template failed to apply the policy. Otherwise 'success'.resultkarmada-controller-manager
binding_sync_work_duration_secondsHistogramDuration in seconds to sync works for a binding object. By the result, 'error' means a binding failed to sync works. Otherwise 'success'.resultkarmada-controller-manager
work_sync_workload_duration_secondsHistogramDuration in seconds to sync the workload to a target cluster. By the result, 'error' means a work failed to sync workloads. Otherwise 'success'.resultkarmada-controller-manager
karmada-agent
policy_preemption_totalCounterNumber of preemption for the resource template. By the result, 'error' means a resource template failed to be preempted by other propagation policies. Otherwise 'success'.resultkarmada-controller-manager
cronfederatedhpa_process_duration_secondsHistogramDuration in seconds to process a CronFederatedHPA. By the result, 'error' means a CronFederatedHPA failed to be processed. Otherwise 'success'.resultkarmada-controller-manager
cronfederatedhpa_rule_process_duration_secondsHistogramDuration in seconds to process a CronFederatedHPA rule. By the result, 'error' means a CronFederatedHPA rule failed to be processed. Otherwise 'success'.resultkarmada-controller-manager
pool_get_operation_totalCounterTotal times of getting from poolname
from
karmada-controller-manager
karmada-agent
pool_put_operation_totalCounterTotal times of putting from poolname
to
karmada-controller-manager
karmada-agent