Global Metrics¶

This is a list of metrics exposed by nais for all applications.

Common Labels¶

The following labels are common to most metrics:

Label	Description
`namespace`	The Kubernetes namespace in which the application is running.
`pod`	The name of the pod in which the application is running. This will be unique for each instance of your application.
`container`	The name of the container in which the application is running. This will be the same as your application name.
`service`	The name of the service in which the application is running.
`node`	The name of the node on which the application is running.
`image`	The name of the container image in which the application is running.

Memory and CPU Usage¶

Memory Usage¶

The container_memory_usage_bytes metric tracks the amount of memory used by a container.

container_memory_usage_bytes{container="my-application", ...}

CPU Usage¶

The container_cpu_usage_seconds_total metric tracks the amount of CPU time used by a container.

container_cpu_usage_seconds_total{container="my-application", ...}

Since this metric is cumulative, you can calculate the rate of CPU usage by using the rate function.

rate(container_cpu_usage_seconds_total{container="my-application", ...}[5m])

Limits and Requests¶

The kube_pod_container_resource_limits and kube_pod_container_resource_requests metrics track the resource limits and requests for a container.

kube_pod_container_resource_limits{resource="memory", container="my-application", ...}
kube_pod_container_resource_requests{resource="memory", container="my-application", ...}
kube_pod_container_resource_limits{resource="cpu", container="my-application", ...}
kube_pod_container_resource_requests{resource="cpu", container="my-application", ...}

Out of Memory (OOMKilled)¶

Out of memory restarts occur when a container exceeds its memory limit and the kernel kills the process to free up memory. OOM kills can be caused by a variety of factors, including memory leaks, excessive memory usage, and insufficient memory limits.

The kube_pod_container_status_terminated_reason metric tracks the number of times a container has been terminated for various reasons, including OOM kills.

kube_pod_container_status_terminated_reason{reason="OOMKilled", ...}

Other reasons for container termination include Error, Completed, and ContainerCannotRun.