Infrastructure metrics

The metrics listed below are provided by the Beats shippers. Each system type requires their corresponding identity field to be in the same event document:

  • Hosts require host.name
  • Docker containers require container.id
  • Kubernetes pods require kubernetes.pod.uid

The event.dataset field is required to display data properly in some views. This field is a combination of metricset.module, which is the Metricbeat module name, and metricset.name, which is the metricset name.

Host Metrics

CPU Usage
Average of system.cpu.user.pct added to the average of system.cpu.system.pct divided by system.cpu.cores
Memory Usage
Average of system.memory.actual.used.pct
Load
Average of system.load.5
Inbound Traffic
Derivative of the maximum of system.network.in.bytes scaled to a 1 second rate
Outbound Traffic
Derivative of the maximum of system.network.out.bytes scaled to a 1 second rate
Log Rate
Derivative of the cumulative sum of the document count scaled to a 1 second rate. This metric relies on the same indices as the logs.

Docker Container Metrics

CPU Usage
Average of docker.cpu.total.pct
Memory Usage
Average of docker.memory.usage.pct
Inbound Traffic
Derivative of the maximum of docker.network.in.bytes scaled to a 1 second rate
Outbound Traffic
Derivative of the maximum of docker.network.out.bytes scaled to a 1 second rate

Kubernetes Pod Metrics

CPU Usage
Average of kubernetes.pod.cpu.usage.node.pct
Memory Usage
Average of kubernetes.pod.memory.usage.node.pct
Inbound Traffic
Derivative of the maximum of kubernetes.pod.network.rx.bytes scaled to a 1 second rate
Outbound Traffic
Derivative of the maximum of kubernetes.pod.network.tx.bytes scaled to a 1 second rate