Skip to content

Latest commit

 

History

History

nrdot-collector-k8s

README.md

nrdot-collector-k8s

Status
Distro nrdot-collector-k8s
Stability preview
Artifacts Docker images on DockerHub

A distribution of the NRDOT collector focused on gathering metrics in a kubernetes environment with two different configs:

  • config-daemonset.yaml (default): Typically deployed as a DaemonSet. Collects node-level metrics via hostmetricsreceiver, filelogreceiver, kubeletstatsreceiver and prometheusreceiver (cAdvisor, kubelet).
  • config-deployment.yaml: Typically deployed as a Deployment. Collects cluster-level metrics via k8seventsreceiver, prometheusreceiver (kube-state-metrics, apiserver, controller-manager, scheduler). Can be enabled by overriding the default docker CMD, i.e. --config /etc/nrdot-collector-k8s/config-deployment.yaml.

Note: See general README for information that applies to all distributions.

Installation

The distribution's main purpose is to be a building block for the nr-k8s-otel-collector helm chart which we recommend using. The helm chart takes care of a lot of configuration required to ensure a smooth operation of the collector and drive the NR Kubernetes experience. If you choose not to use the helm chart, you'll have to follow the general installation and provide the necessary permissions for the collector to access the k8s APIs yourself, see also our troubleshooting guide.

Dependencies

  • Most k8s APIs scraped by the various receivers require additional permissions setup which are provided by the nr-k8s-otel-collector chart out of the box in the form of a service account. If you wish to add those permissions by hand, please refer to the chart itself.
  • kube-state-metrics is required to be running in your cluster. The metrics emitted by this add-on are used to create NR entities for various k8s resources.

Configuration

If you use the nr-k8s-otel-collector helm chart, please refer to its documentation for configuration options.

See general README for information that applies to all distributions.

Distribution-specific configuration

Environment Variable Description Default
K8S_CLUSTER_NAME Kubernetes Cluster Name used to populate attributes like k8s.cluster.name cluster-name-placeholder
MY_POD_IP Pod IP to configure otlpreceiver cluster-name-placeholder

Limitations of the standalone image (vs helm chart)

Key differences are formally documented in the script that ensures the configurations stay in-sync and can be summarized as:

  • lowDataMode: false is hardcoded and cannot be toggled easily (it is always possible to overwrite the config with native collector options)
  • hostmetricsreceiver in the daemonset config does not assume that the host file system is mounted, thus providing metrics about its container and not the host node.
  • healthcheckextension is enabled by default