Network Performance Monitoring is now generally available! Network Monitoring is now available!

Kubernetes Host Setup

Installing the Agent directly on your host (rather than having the Agent run in a Pod) provides additional visibility into your ecosystem, independent of Kubernetes.

To gather your kube-state metrics:

  1. Download the Kube-State manifests folder.

  2. Apply them to your Kubernetes cluster:


Setup the kube-dns integration


Since Agent v6, Kubernetes DNS integration works automatically with the Autodiscovery.

Note: these metrics are unavailable for Azure Kubernetes Service (AKS).

Collect container logs

Available for Agent >6.0

There are two ways to collect logs from containers running in Kubernetes:

  • Through the Docker socket.
  • Through the Kubernetes log files.

Datadog recommends using the Kubernetes log files approach when you are either not using Docker, or are using more than 10 containers per pod.

Datadog also recommends that you take advantage of DaemonSets to automatically deploy the Datadog Agent on all your nodes. Otherwise, to manually enable log collection from one specific node, add the following parameters in the datadog.yaml:

logs_enabled: true
  - name: kubelet
  - name: kubelet
    polling: true
  container_collect_all: true

Restart the Agent.

Use Autodiscovery with Pod Annotations to configure log collection to add multiline processing rules, or to customize the source and service attributes.

Further Reading

To get a better idea of how (or why) to integrate your Kubernetes service, see the related series of Datadog blog posts.