Try it free

Get started with Kubernetes platform monitoring + Application observability

  • Latest Dynatrace
  • Getting started guide
  • 8-min read
  • Published Aug 31, 2026

Deploy Dynatrace Operator to monitor your Kubernetes cluster and the applications running on it. Application monitoring is licensed per container-hour as Container Application Monitoring.

Explore use cases

With Kubernetes platform monitoring + Application observability deployed, you can use the Kubernetes (new) Kubernetes app to:

  • Assess and troubleshoot Kubernetes cluster and workload health
  • Optimize Kubernetes workload resource usage
  • Receive alerts and events to detect and respond to cluster anomalies
  • Explore metrics, events, and logs for your pods and nodes in a single interface
  • Troubleshoot common health problems of Kubernetes workloads
  • Automatic distributed tracing across containers
  • Code-level and service insights in application containers

If you’re looking to gain deeper insights into your Kubernetes environment, check out the deployment overview to learn about Full-Stack monitoring, which adds host and VM visibility. You can further extend with Log Analytics, Code Monitoring, Real User Monitoring and Synthetics, and Application Security.

Who is this for?

This tutorial is intended for platform engineers, site reliability engineers (SREs), and Kubernetes administrators.

What will you learn?

By the end of this tutorial, you'll have Dynatrace monitoring both your Kubernetes cluster and the applications running on it. Along the way, you'll:

  • Install Dynatrace Operator via Helm or manifest.
  • Configure the DynaKube custom resource for Container Application Monitoring.
  • Verify your deployment.

Before you begin

Before installing Dynatrace on your Kubernetes cluster, ensure that you meet the following requirements:

  • Your kubectl CLI is connected to the Kubernetes cluster that you want to monitor.
  • You have sufficient privileges on the monitored cluster to run kubectl or oc commands. If you don't use the cluster-admin cluster role, see deployment permissions for the required permissions.

Cluster setup and configuration

  • You must allow egress for Dynatrace pods (default: Dynatrace namespace) to your Dynatrace environment URL.

  • For OpenShift Dedicated, you need the cluster-admin role.

  • Helm installation Use Helm version 3.

Supported versions

See supported Kubernetes/OpenShift platform versions and distributions.

Container Application Monitoring works on all supported distributions. For setups without Dynatrace Operator and for other platforms, see Monitor container workloads.

If you're deploying to OpenShift, configure SCC before proceeding; it's required for applicationMonitoring deployments with the CSI driver.

Do not combine with host monitoring

Don't combine hostMonitoring and applicationMonitoring in the same cluster. Container Application Monitoring needs application monitoring mode: in a host monitoring mode, OneAgent already licenses each node as a host, so its containers would be licensed twice.

Licensing

Dynatrace Operator version 1.10+ OneAgent version 1.345+

Kubernetes platform monitoring + Application observability requires a Dynatrace Platform Subscription (DPS) (SaaS).

  • Kubernetes Platform Monitoring is licensed by number of pods per hour (pod-hours).
  • Container Application Monitoring is licensed by number of containers per hour (container-hours).

For more information on included trace and metric volumes per container-hour, see Container Application Monitoring consumption.

Code Monitoring is licensed separately, and Profiling isn't available. For all limitations, see Monitor container workloads. To attribute costs to teams, see cost allocation.

Get started

Select the installation method that works best for you: Helm or manifest.

Install with Helm Helm

Dynatrace Operator version 1.10+

These instructions use Dynatrace's Helm chart from an OCI registry. If you have the Dynatrace repository in your local Helm repositories, you can remove it before proceeding:

helm repo remove dynatrace
  1. Install Dynatrace Operator

    If you're using Helm 4.0+, replace --atomic with --rollback-on-failure in the commands below.

    You have two options:

    Default installation / OCI registry installation

    The following command works for both default installations and installations using an OCI registry.

    helm install dynatrace-operator oci://public.ecr.aws/dynatrace/dynatrace-operator \
    --create-namespace \
    --namespace dynatrace \
    --atomic
    Installation with additional configuration of the Helm chart

    Edit the values.yaml sample from GitHub, and then run the install command, passing the YAML file as an argument:

    helm install dynatrace-operator oci://public.ecr.aws/dynatrace/dynatrace-operator \
    --create-namespace \
    --namespace dynatrace \
    --atomic \
    -f values.yaml

    If installCRD is set to false, you need to create the custom resource definition manually before starting the Helm installation:

    kubectl apply -f https://github.com/Dynatrace/dynatrace-operator/releases/download/v1.10.2/dynatrace-operator-crd.yaml

    VMware Tanzu Kubernetes (TKGI) and IBM Kubernetes Service (IKS) require additional configuration.

  2. Create a secret named dynakube for the Dynatrace Operator token and data ingest token obtained in Tokens and permissions required.

    kubectl -n dynatrace create secret generic dynakube \
    --from-literal="apiToken=<OPERATOR_TOKEN>" \
    --from-literal="dataIngestToken=<DATA_INGEST_TOKEN>"
  3. Create your DynaKube custom resource YAML file.

    When do you need the opt-in attribute?

    For OneAgent workloads, you need the resource attribute only if the tenant-wide setting builtin:billing.container-application-monitoring.optin isn't enabled. It's enabled by default for new customers, so they can leave the attribute out of the DynaKube.

    For OpenTelemetry workloads, you always need the resource attribute. The tenant-wide setting covers OneAgent only.

    For all opt-in options, see Kubernetes container workloads.

    DynaKube custom resource sample—OneAgent (Container Application Monitoring)

    The following sample enables Container Application Monitoring for OneAgent-instrumented containers in namespaces labeled container_application_monitoring-oneagent: "true".

    You can review the available parameters or how-to guides and adapt the DynaKube according to your requirements.

    apiVersion: dynatrace.com/v1beta6
    kind: DynaKube
    metadata:
    name: dynakube
    namespace: dynatrace
    spec:
    apiUrl: https://<ENVIRONMENTID>.live.dynatrace.com/api
    tokens: dynakube
    # Opt in to Container Application Monitoring billing.
    # Omit this block if the tenant-wide setting
    # builtin:billing.container-application-monitoring.optin is already enabled.
    resourceAttributes:
    dt.has_container_application_monitoring: "true"
    oneAgent:
    applicationMonitoring:
    namespaceSelector:
    matchLabels:
    container_application_monitoring-oneagent: "true" # Adjust to target your namespaces.
    activeGate:
    capabilities:
    - routing
    - kubernetes-monitoring
    resources:
    requests:
    cpu: 500m
    memory: 512Mi
    limits:
    cpu: 1000m
    memory: 1.5Gi
    DynaKube custom resource sample—OpenTelemetry (Container Application Monitoring)

    In the following sample, the Operator injects the OTLP exporter settings and the opt-in attribute into pods in namespaces labeled container_application_monitoring-otlp: "true". It also adds k8s.pod.uid and k8s.container.name, which Dynatrace needs to attribute telemetry to a container.

    apiVersion: dynatrace.com/v1beta6
    kind: DynaKube
    metadata:
    name: dynakube
    namespace: dynatrace
    spec:
    apiUrl: https://<ENVIRONMENTID>.live.dynatrace.com/api
    tokens: dynakube
    # Opt in to Container Application Monitoring billing (applies to all components).
    # Required for OTLP workloads regardless of the tenant-wide setting.
    resourceAttributes:
    dt.has_container_application_monitoring: "true"
    # The Operator injects OTLP exporter configuration (endpoint, token,
    # resource attributes) into pods in the selected namespaces.
    otlpExporterConfiguration:
    signals:
    traces: {}
    metrics: {}
    namespaceSelector:
    matchLabels:
    container_application_monitoring-otlp: "true" # Adjust to target your namespaces.
    activeGate:
    capabilities:
    - routing
    - kubernetes-monitoring
    resources:
    requests:
    cpu: 500m
    memory: 512Mi
    limits:
    cpu: 1000m
    memory: 1.5Gi

    To opt in individual workloads, OpenTelemetry workloads the Operator doesn't manage, or sidecar containers, see Kubernetes container workloads.

  4. Optional Monitor potentially sensitive data

    To monitor potentially sensitive data such as Secrets (values are masked before ingest) and ConfigMaps, you need to create a ClusterRole that allows access to these resources. Add the following snippet at the end of your DynaKube custom resource YAML file.

    Installation with sensitive data monitoring
    ---
    apiVersion: rbac.authorization.k8s.io/v1
    kind: ClusterRole
    metadata:
    labels:
    rbac.dynatrace.com/aggregate-to-monitoring: "true"
    name: dynatrace-kubernetes-monitoring-sensitive
    rules:
    - apiGroups:
    - ""
    resources:
    - configmaps
    - secrets
    verbs:
    - list
    - watch
    - get

    In Dynatrace Operator version 1.8.x, the role is automatically bound to the dynatrace-kubernetes-monitoring ClusterRole via aggregation rules. For more information, see ClusterRole aggregation documentation. Starting with Dynatrace Operator version 1.9.0, aggregated ClusterRoles are no longer used. Create a ClusterRoleBinding to bind this ClusterRole to the dynatrace-activegate service account directly.

  5. Apply the DynaKube custom resource

    Run the command below to apply the DynaKube custom resource, replacing <your-DynaKube-CR> with your file name. A validation webhook will provide helpful error messages if there's a problem.

    kubectl apply -f <your-DynaKube-CR>.yaml
  6. Optional Verify deployment

    Verify that your DynaKube is running and all pods in your Dynatrace namespace are running and ready.

    > kubectl get dynakube -n dynatrace
    NAME APIURL STATUS AGE
    dynakube https://<ENVIRONMENTID>.live.dynatrace.com/api Running 45s

    In a default DynaKube configuration with CSI driver, you should see the following pods:

    > kubectl get pods -n dynatrace
    NAME READY STATUS RESTARTS AGE
    dynakube-activegate-0 1/1 Running 0 50s
    dynatrace-oneagent-csi-driver-qxfwx 4/4 Running 0 2m49s
    dynatrace-oneagent-csi-driver-xk5c4 4/4 Running 0 2m49s
    dynatrace-oneagent-csi-driver-mz6ch 4/4 Running 0 2m49s
    dynatrace-operator-7dc8dc7d8c-wmh4z 1/1 Running 0 2m59s
    dynatrace-webhook-7bb6957fb5-l8fsq 1/1 Running 0 2m59s
    dynatrace-webhook-7bb6957fb5-rqnqk 1/1 Running 0 2m59s

Install with manifest

Kubernetes
  1. Create a dynatrace namespace

    kubectl create namespace dynatrace
  2. Install Dynatrace Operator

    kubectl apply -f https://github.com/Dynatrace/dynatrace-operator/releases/download/v1.10.2/kubernetes-csi.yaml
    Without CSI driver
    kubectl apply -f https://github.com/Dynatrace/dynatrace-operator/releases/download/v1.10.2/kubernetes.yaml

    VMware Tanzu Kubernetes (TKGI) and IBM Kubernetes Service (IKS) require additional configuration.

    Run the following command to see when Dynatrace Operator components finish initialization:

    kubectl -n dynatrace wait pod --for=condition=ready \
    --selector=app.kubernetes.io/name=dynatrace-operator,app.kubernetes.io/component=webhook \
    --timeout=300s
  3. Create a secret named dynakube for the Dynatrace Operator token and data ingest token.

    kubectl -n dynatrace create secret generic dynakube \
    --from-literal="apiToken=<OPERATOR_TOKEN>" \
    --from-literal="dataIngestToken=<DATA_INGEST_TOKEN>"
  4. Create your DynaKube custom resource YAML file.

    See the DynaKube samples in the Helm section above for reference configurations for OneAgent and OpenTelemetry deployments. You can review available parameters and how-to guides and adapt the DynaKube to your requirements.

  5. Optional Monitor potentially sensitive data

    To monitor potentially sensitive data such as Secrets (values are masked before ingest) and ConfigMaps, you need to create a ClusterRole that allows access to these resources. Add the following snippet at the end of your DynaKube custom resource YAML file.

    Installation with sensitive data monitoring
    ---
    apiVersion: rbac.authorization.k8s.io/v1
    kind: ClusterRole
    metadata:
    labels:
    rbac.dynatrace.com/aggregate-to-monitoring: "true"
    name: dynatrace-kubernetes-monitoring-sensitive
    rules:
    - apiGroups:
    - ""
    resources:
    - configmaps
    - secrets
    verbs:
    - list
    - watch
    - get

    In Dynatrace Operator version 1.8.x, the role is automatically bound to the dynatrace-kubernetes-monitoring ClusterRole via aggregation rules. For more information, see ClusterRole aggregation documentation. Starting with Dynatrace Operator version 1.9.0, aggregated ClusterRoles are no longer used. Create a ClusterRoleBinding to bind this ClusterRole to the dynatrace-activegate service account directly.

  6. Apply the DynaKube custom resource

    kubectl apply -f <your-DynaKube-CR>.yaml
  7. Optional Verify deployment

    > kubectl get dynakube -n dynatrace
    NAME APIURL STATUS AGE
    dynakube https://<ENVIRONMENTID>.live.dynatrace.com/api Running 45s
OpenShift
  1. Add a dynatrace project

    oc adm new-project --node-selector="" dynatrace
  2. Install Dynatrace Operator

    oc apply -f https://github.com/Dynatrace/dynatrace-operator/releases/download/v1.10.2/openshift-csi.yaml
    Without CSI driver
    oc apply -f https://github.com/Dynatrace/dynatrace-operator/releases/download/v1.10.2/openshift.yaml
    oc -n dynatrace wait pod --for=condition=ready \
    --selector=app.kubernetes.io/name=dynatrace-operator,app.kubernetes.io/component=webhook \
    --timeout=300s
  3. Create a secret named dynakube.

    oc -n dynatrace create secret generic dynakube \
    --from-literal="apiToken=<OPERATOR_TOKEN>" \
    --from-literal="dataIngestToken=<DATA_INGEST_TOKEN>"
  4. Create your DynaKube custom resource YAML file (see the Helm section above for reference samples) and apply it.

    oc apply -f <your-DynaKube-CR>.yaml
  5. Optional Verify deployment

    > oc get dynakube -n dynatrace
    NAME APIURL STATUS AGE
    dynakube https://<ENVIRONMENTID>.live.dynatrace.com/api Running 45s

Congratulations!

You've successfully set up Kubernetes platform monitoring and Application observability, billed as Container Application Monitoring. In this tutorial, you:

  • Installed Dynatrace Operator (with optional CSI driver) on your cluster.
  • Configured and applied a DynaKube custom resource with the Container Application Monitoring opt-in.

Next steps

Kubernetes

Monitor and optimize Kubernetes with Dynatrace. Get real-time insights and health for clusters and workloads.

Kubernetes

Guides

Detailed description of installation and configuration options for specific use-cases

Guides

Troubleshooting

This page will assist you in navigating any challenges you may encounter while working with the Dynatrace Operator and its various components.

Troubleshooting

How it works

In-depth description on how the deployment on Kubernetes works.

How it works

Reference

Contains a reference page with configuration options for each Dynatrace component

Reference

Dynatrace Operator release notes

Release notes for Dynatrace Operator

Dynatrace Operator release notes

Update or uninstall Dynatrace Operator

Upgrade paths, update procedures, and uninstallation guide for Dynatrace Operator.

Update or uninstall Dynatrace Operator

Size Dynatrace ActiveGates in Kubernetes

Find CPU and memory resource recommendations for Dynatrace ActiveGates deployed in Kubernetes, sized by cluster scale and workload type.

Size Dynatrace ActiveGates in Kubernetes

Related topics

  • Kubernetes Classic
  • Understand and manage Container Application Monitoring consumption (DPS)