Deploying Datadog in Kubernetes using Helm

Helm is a package management tool for Kubernetes.

Installing Helm

Installing the Helm client

brew install kubernetes-helm

sudo snap install helm --classic

choco install kubernetes-helm

For other platforms and methods of installing Helm, refer to the Helm documentation.

Installing the Helm server (Tiller)

Note: This is not required for versions of Helm greater than 3.0.0. Skip to Installing the Datadog Helm chart if this applies to you.

If your Kubernetes environment does not use RBAC, the following command installs Tiller in your cluster:

helm init

Refer to Helm’s Tiller documentation for further details.

If your Kubernetes cluster is RBAC-enabled, use the following RBAC to deploy Tiller.

apiVersion: v1
kind: ServiceAccount
metadata:
 name: tiller
 namespace: kube-system
---
apiVersion: rbac.authorization.k8s.io/v1beta1
kind: ClusterRoleBinding
metadata:
 name: tiller
roleRef:
 apiGroup: rbac.authorization.k8s.io
 kind: ClusterRole
 name: cluster-admin
subjects:
 - kind: ServiceAccount
   name: tiller
   namespace: kube-system

With this as your tiller-rbac-config.yaml, then run:

kubectl create -f tiller-rbac-config.yaml

helm init --service-account tiller

Refer to Helm’s Tiller/RBAC documentation for further details.

Verify installation

To verify your installation, run:

kubectl get pods -n kube-system -l app=helm

This returns something resembling:

NAME READY STATUS RESTARTS AGE
tiller-deploy-f54b67464-jl5gm 1/1 Running 0 3h16m

Installing the Datadog Helm chart

To install the chart with the release name <RELEASE_NAME>, retrieve your Datadog API key from your Agent installation instructions and run:

helm install --name <RELEASE_NAME> --set datadog.apiKey=<DATADOG_API_KEY> stable/datadog

helm install <RELEASE_NAME> --set datadog.apiKey=<DATADOG_API_KEY> stable/datadog

This chart adds the Datadog Agent to all nodes in your cluster via a DaemonSet. It also optionally deploys the kube-state-metrics chart and uses it as an additional source of metrics about the cluster. A few minutes after installation, Datadog begins to report hosts and metrics.

Note: For a full list of the Datadog chart’s configurable parameters and their default values, refer to the Datadog Helm repository README.

Configuring the Datadog Helm chart

As a best practice, a YAML file that specifies the values for the chart parameters should be provided to configure the chart:

1. Copy the default datadog-values.yaml value file.
2. Set the apiKey parameter with your Datadog API key.

3. Upgrade the Datadog Helm chart with the new datadog-values.yaml file:

   helm upgrade -f datadog-values.yaml <RELEASE_NAME> stable/datadog --recreate-pods

Enable Log Collection

Update your datadog-values.yaml file with the following log collection configuration, then upgrade your Datadog Helm chart:

datadog:
  (...)
 logsEnabled: true
 logsConfigContainerCollectAll: true

Enable Custom Metrics Collection

To gather custom metrics with DogStatsD, update your datadog-values.yaml file to enable non-local traffic.

datadog:
  (...)
  nonLocalTraffic: true

Enable APM and Distributed Tracing

Note: If you want to deploy the Datadog Agent as a deployment instead of a DaemonSet, configuration of APM via Helm is not supported.

Update your datadog-values.yaml file with the following APM configuration:

datadog:
  (...)
  apmEnabled: true
  nonLocalTraffic: true

(...)

daemonset:
  (...)
  useHostPort: true

Update the env section of your application’s manifest with the following:

env:
  - name: DD_AGENT_HOST
    valueFrom:
      fieldRef:
        fieldPath: status.hostIP

Then upgrade your Datadog Helm chart.

Finally, point your application-level tracers to the host IP using the environment variable DD_AGENT_HOST. For example, in Python:

import os
from ddtrace import tracer

tracer.configure(
    hostname=os.environ['DD_AGENT_HOST'],
    port=os.environ['DD_TRACE_AGENT_PORT'],
)

Refer to the language-specific APM instrumentation docs for more examples.

Enabling Process Collection

Update your datadog-values.yaml file with the following process collection configuration, then upgrade your Datadog Helm chart:

datadog:
  (...)
  processAgentEnabled: true

Enabling integrations with Helm

The Datadog entrypoint copies files with a .yaml extension found in /conf.d and files with .py extension in /check.d to /etc/datadog-agent/conf.d and /etc/datadog-agent/checks.d respectively. The keys for datadog.confd and datadog.checksd should mirror the content found in their respective ConfigMaps, i.e.:

datadog:
  confd:
    redisdb.yaml: |-
      ad_identifiers:
        - redis
        - bitnami/redis
      init_config:
      instances:
        - host: "%%host%%"
          port: "%%port%%"
    jmx.yaml: |-
      ad_identifiers:
        - openjdk
      instance_config:
      instances:
        - host: "%%host%%"
          port: "%%port_0%%"

Uninstalling the Datadog Helm Chart

To uninstall or delete a deployment called <RELEASE_NAME>:

helm delete <RELEASE_NAME>

This command removes all Kubernetes components associated with the chart and deletes the release.