Datadog on Kubernetes

1. Architecture #

The components.

Components #

2. Installing with the Operator #

Declarative.

Installation #

helm repo add datadog https://helm.datadoghq.com && helm repo update
helm install datadog-operator datadog/datadog-operator
kubectl create secret generic datadog-secret --from-literal api-key="$DD_API_KEY"

apiVersion: datadoghq.com/v2alpha1
kind: DatadogAgent
metadata:
  name: datadog
spec:
  global:
    clusterName: prod-eu
    site: datadoghq.eu
    credentials:
      apiSecret: { secretName: datadog-secret, keyName: api-key }
  features:
    logCollection: { enabled: true, containerCollectAll: true }
    apm:
      instrumentation: { enabled: true }      # injection automatique des bibliothèques APM
    liveProcessCollection: { enabled: true }
    orchestratorExplorer: { enabled: true }

Good to know: The datadog/datadog Helm chart remains an alternative; the Operator has the advantage of a single resource, validated and easy to version with GitOps.

3. Autodiscovery #

Monitoring whatever arrives.

Annotations #

Autodiscovery applies a check configuration to each matching container as soon as it starts. The configuration is set as annotations on the pod:

metadata:
  annotations:
    ad.datadoghq.com/redis.checks: |
      { "redisdb": { "instances": [ { "host": "%%host%%", "port": "6379" } ] } }
    ad.datadoghq.com/redis.logs: '[{"source": "redis", "service": "cache"}]'
  labels:
    tags.datadoghq.com/env: prod
    tags.datadoghq.com/service: cache
    tags.datadoghq.com/version: "7.2"

Good practice: The tags.datadoghq.com/* labels apply unified tagging to the pod's metrics, logs and traces: it is the first thing to standardize in your charts.

4. Troubleshooting #

Useful commands.

Commands #

kubectl exec -it ds/datadog-agent -- agent status          # état des checks, logs, APM
kubectl exec -it ds/datadog-agent -- agent configcheck     # configurations Autodiscovery résolues
kubectl exec -it deploy/datadog-cluster-agent -- datadog-cluster-agent status