Yandex Managed Service for Kubernetes metrics

Written by

Updated at June 25, 2024

This section describes the Managed Service for Kubernetes metrics delivered to Monitoring.

The metric name is written in the name label.

All Managed Service for Kubernetes metrics share the service=managed-kubernetes label.

Container metrics

Common labels for all container metrics:

Container metrics:

Name Type, unit	Description
`container.cpu.core_usage_time` `COUNTER`, milliseconds	Total CPU usage on all cores used by the container
`container.cpu.limit_cores` `DGAUGE`, %	Limit on the number of CPU cores per container
`container.cpu.limit_utilization` `DGAUGE`, %	CPU core utilization
`container.cpu.request_cores` `DGAUGE`, %	Number of CPU cores requested by the container
`container.cpu.request_utilization` `DGAUGE`, %	Utilization of CPU cores requested by the container
`container.ephemeral_storage.used_bytes` `IGAUGE`, bytes	Ephemeral storage usage
`container.memory.limit_bytes` `IGAUGE`, bytes	Limit on the amount of container memory
`container.memory.limit_utilization` `DGAUGE`, %	Utilized container memory
`container.memory.page_fault_count` `COUNTER`, bytes	Memory page errors by type: major and minor. Additional labels: `fault_type`.
`container.memory.request_bytes` `IGAUGE`, bytes	Amount of memory requested by the container
`container.memory.request_utilization` `DGAUGE`, %	Utilization of memory requested by the container
`container.memory.used_bytes` `IGAUGE`, bytes	Memory usage by the container
`container.memory.working_set_bytes` `IGAUGE`, bytes	Container working set memory used
`container.restart_count` `COUNTER`, number	Number of container restarts
`container.uptime_seconds` `DGAUGE`, milliseconds	Container uptime since its start

Common labels for all master metrics:

Label	Value
cluster_id	Cluster ID
host	Cluster ID

Master metrics:

Name Type, unit	Description
`master.cpu.utilization_percent` `DGAUGE`, %	CPU utilization
`master.memory.utilization_percent` `DGAUGE`, %	Utilized memory

Common labels for all node metrics:

Label	Value
cluster_id	Cluster ID
node	Node name

Node metrics:

Name Type, unit	Description
`node.cpu.allocatable_cores` `DGAUGE`, %	Amount of CPU resources available on the node
`node.cpu.core_usage_time` `COUNTER`, milliseconds	Total CPU usage on all cores used by the node
`node.cpu.total_cores` `DGAUGE`, number	Total number of node processor cores
`node.ephemeral_storage.allocatable_bytes` `IGAUGE`, bytes	Amount of ephemeral storage resources available on the node
`node.ephemeral_storage.inodes_free` `IGAUGE`, number	Number of free index descriptors
`node.ephemeral_storage.inodes_total` `IGAUGE`, number	Total number of index descriptors
`node.ephemeral_storage.total_bytes` `IGAUGE`, bytes	Ephemeral storage size
`node.ephemeral_storage.used_bytes` `IGAUGE`, bytes	Ephemeral storage used by the node
`node.memory.allocatable_bytes` `IGAUGE`, bytes	Amount of RAM resources available on the node
`node.memory.total_bytes` `IGAUGE`, bytes	Total amount of node memory
`node.memory.used_bytes` `IGAUGE`, bytes	Memory used by the node
`node.memory.working_set_bytes` `IGAUGE`, bytes	Node working set memory used
`node.network.received_bytes_count` `COUNTER`, bytes	Incoming traffic (bytes received from the network)
`node.network.sent_bytes_count` `COUNTER`, bytes	Outgoing traffic (bytes sent to the network)
`node.pid.limit` `IGAUGE`, number	Limit on the number of process IDs per node
`node.pid.used` `IGAUGE`, number	Process IDs used in the node

Common labels for all pod metrics:

Pod metrics:

Name Type, unit	Description
`pod.cpu.core_usage_time` `COUNTER`, milliseconds	Total CPU usage on all cores used by the pod
`pod.ephemeral_storage.used_bytes` `IGAUGE`, bytes	Ephemeral storage used by the pod
`pod.memory.used_bytes` `IGAUGE`, bytes	Memory used by the pod
`pod.memory.working_set_bytes` `IGAUGE`, bytes	Pod working set memory used
`pod.network.received_bytes_count` `COUNTER`, bytes	Incoming traffic (bytes received from the network)
`pod.network.sent_bytes_count` `COUNTER`, bytes	Outgoing traffic (bytes sent to the network)
`pod.volume.total_bytes` `IGAUGE`, bytes	Total size of temporary volume. Additional labels: `volume`.
`pod.volume.used_bytes` `IGAUGE`, bytes	Temporary volume usage. Additional labels: `volume`.
`pod.volume.utilization` `DGAUGE`, %	Temporary volume utilization. Additional labels: `volume`.

Common labels for all persistent volume metrics:

Label	Value
cluster_id	Cluster ID
namespace	Namespace name
persistentvolumeclaim	Persistent volume claim name

Persistent volume metrics:

Name Type, unit	Description
`persistent_volume.inodes_free` `IGAUGE`, number	Number of free index descriptors
`persistent_volume.inodes_total` `IGAUGEE`, number	Total number of index descriptors
`persistent_volume.total_bytes` `IGAUGE`, bytes	Total size of the persistent volume
`persistent_volume.used_bytes` `IGAUGE`, bytes	Persistent volume usage
`persistent_volume.utilization` `DGAUGE`, %	Persistent volume utilization