> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/resilience-testing/chaos-engineering/use-chaos-engineering/infrastructure/types/ddcr/installation.md).
# Installation
Steps to install Harness Delegate and enable chaos.
This topic describes how you can install Harness Delegate and enable chaos on it. You can follow the interactive demo or the steps to install Harness Delegate.
{% hint style="info" %}
**DELEGATE IMAGE OPTIONS**
Harness provides two delegate image types:
* **Standard Delegate Image** (Recommended) - Includes all required binaries for chaos experimentation and discovery. Follow the standard installation steps below.
* **Delegate-Minimal Image** - Lightweight image that requires additional configuration. See [Using Delegate-Minimal Image](#using-delegate-minimal-image) if you specifically need to use the minimal image.
{% endhint %}
### Standard Delegate Installation
This section covers the installation of the standard delegate image.
#### Prerequisites
* [Create an Environment](/resilience-testing/chaos-engineering/new-to-chaos-engineering/on-premise-vs-saas.md)
#### Interactive Demo
Navigate through the demo below to see how to create an environment, a new infrastructure, a new connector and a Delegate.
If you have created a connector and a Harness Delegate earlier, you can find their names in the list that is displayed while selecting the connector and Delegate, respectively. Select it instead of creating a new one.
{% embed url="" %}
#### Step-by-Step Installation
1. Select the **Environment** you created, and choose the **Infrastructure type**. In this example, select **Kubernetes** and click **+ New Infrastructure**.
2. To use Harness Delegate, select **Yes**. To use a dedicated chaos infrastructure, select **No**. In this example, select **Yes**, and click **Next**.
3. Provide a **Name**, **Deployment Type**, **Select Infrastructure Type**, and provide **Cluster Details**, such as **Connector**, and **Namespace**.
a. To create a new connector, go to [Create Kubernetes Connector](/harness-ai/use-harness-platform/connectors/cloud-providers/add-a-kubernetes-cluster-connector.md#add-a-kubernetes-cluster-connector). Harness CE supports creating two types of connectors:
* [Kubernetes (Direct Connection)](/harness-ai/use-harness-platform/connectors/cloud-providers/add-a-kubernetes-cluster-connector/kubernetes-cluster-connector-settings-reference.md)
* [GCP](/harness-ai/use-harness-platform/connectors/cloud-providers/connect-to-google-cloud-platform-gcp/gcs-connector-settings-reference.md)
b. While creating a Kubernetes connector, select or [create a new Delegate](/harness-ai/use-harness-platform/delegates/delegate/install-delegates/overview.md#install-the-helm-chart). Once your Delegate shows up in the list, enter its name in the search bar, and select **Save and Continue**. This will test your connection and describe if it was a success or no. Contact [Harness Support](mailto:support@harness.io) if you encounter issues with the verification.
4. To **Enable Chaos** and use **Configure Advanced Settings (Optional)**, navigate to **Environments** and **Enable Chaos** on the infrastructure you created earlier.
5. Choose between **Beginner** or **Expert** and click **Go**.
6. You can choose to provide additional configurations in the **Configure Advanced Settings (Optional)** section to configure [chaos runner and Discovery Agent](/harness-ai/use-harness-platform/service-discovery/customize-agent.md) and click **Go**. This creates application maps and chaos experiments for you.
{% hint style="info" %}
* Ensure your Delegate version is `24.09.83900` or above to support executing chaos experiments using DDCR.
* Currently, DDCR (or Harness Delegate) can only be installed to execute Kubernetes-based faults. This means you can only create Kubernetes connectors.
* To execute faults on Linux and Windows, install Linux and Windows agents respectively.
* To configure mTLS with DDCR and Discovery Agent, go to [mTLS Support](/resilience-testing/chaos-engineering/use-chaos-engineering/infrastructure/types/ddcr/mtls-support.md).
* To configure mTLS with DDCR and Discovery Agent proxy settings, go to [proxy Support](/resilience-testing/chaos-engineering/use-chaos-engineering/infrastructure/types/ddcr/proxy-support.md).
{% endhint %}
***
### Using Delegate-Minimal Image
{% hint style="warning" %}
**IMPORTANT**
The **standard delegate image** includes all required binaries for chaos experimentation and discovery out-of-the-box. Use the steps above for standard delegate installation.
If you are using the **delegate-minimal image**, additional setup is required as described below.
{% endhint %}
By default, chaos experimentation and discovery are **not supported** with the `delegate-minimal` image. The delegate-minimal image does not include `kubectl` and `go-template` binaries, which are required for discovery and chaos experimentation.
#### Prerequisites for Delegate-Minimal
If you are using a `delegate-minimal` image, you must install `kubectl` and `go-template` binaries using one of the following methods:
#### Option 1: Create a Custom Delegate Image
Create your own custom version of the `delegate-minimal` image with `kubectl` and `go-template` binaries pre-installed.
#### Option 2: Use INIT\_SCRIPT Environment Variable (Recommended)
Provide an installation script in the `INIT_SCRIPT` environment variable of the delegate deployment:
```yaml
- env:
- name: INIT_SCRIPT
value: |
##kubectl
curl -L0 https://app.harness.io/public/shared/tools/kubectl/release/v1.28.7/bin/linux/amd64/kubectl -o kubectl
chmod +x ./kubectl
mv kubectl /usr/local/bin/
## go-template
curl -L0 https://app.harness.io/public/shared/tools/go-template/release/v0.4.1/bin/linux/amd64/go-template -o go-template
chmod +x ./go-template
mv go-template /usr/local/bin/
```
View Complete Deployment Example
Below is a complete example of a delegate deployment using the minimal image with the `INIT_SCRIPT` configuration:
```yaml
apiVersion: apps/v1
kind: Deployment
metadata:
annotations:
deployment.kubernetes.io/revision: "12"
meta.helm.sh/release-name: minimal-delegate
meta.helm.sh/release-namespace: harness-delegate-ng
creationTimestamp: "2025-12-24T14:59:12Z"
generation: 14
labels:
app.kubernetes.io/instance: minimal-delegate
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/name: minimal-delegate
harness.io/name: minimal-delegate
helm.sh/chart: harness-delegate-ng-1.0.30
name: minimal-delegate
namespace: harness-delegate-ng
resourceVersion: "1766592671187615000"
uid: e722c89a-b32a-4967-b0d1-18b30875ec5a
spec:
minReadySeconds: 120
progressDeadlineSeconds: 600
replicas: 1
revisionHistoryLimit: 10
selector:
matchLabels:
app.kubernetes.io/instance: minimal-delegate
app.kubernetes.io/name: minimal-delegate
harness.io/name: minimal-delegate
strategy:
rollingUpdate:
maxSurge: 25%
maxUnavailable: 25%
type: RollingUpdate
template:
metadata:
annotations:
checksum/configmap:
checksum/secret:
prometheus.io/path: /api/metrics
prometheus.io/port: "3460"
prometheus.io/scrape: "true"
creationTimestamp: null
labels:
app.kubernetes.io/instance: minimal-delegate
app.kubernetes.io/name: minimal-delegate
harness.io/name: minimal-delegate
spec:
containers:
- env:
- name: INIT_SCRIPT
value: |
##kubectl
curl -L0 https://app.harness.io/public/shared/tools/kubectl/release/v1.28.7/bin/linux/amd64/kubectl -o kubectl
chmod +x ./kubectl
mv kubectl /usr/local/bin/
## go-template
curl -L0 https://app.harness.io/public/shared/tools/go-template/release/v0.4.1/bin/linux/amd64/go-template -o go-template
chmod +x ./go-template
mv go-template /usr/local/bin/
envFrom:
- configMapRef:
name: minimal-delegate
- secretRef:
name: minimal-delegate
- configMapRef:
name: minimal-delegate-proxy
optional: true
- secretRef:
name: minimal-delegate-proxy
optional: true
- configMapRef:
name: minimal-delegate-shared-certificates
optional: true
image: us-docker.pkg.dev/gar-prod-setup/harness-public/harness/delegate:25.08.86602.minimal
imagePullPolicy: Always
livenessProbe:
failureThreshold: 3
httpGet:
path: /api/health
port: 3460
scheme: HTTP
initialDelaySeconds: 30
periodSeconds: 20
successThreshold: 1
timeoutSeconds: 1
name: delegate
ports:
- containerPort: 8080
protocol: TCP
- containerPort: 3460
name: api
protocol: TCP
resources: {}
securityContext:
allowPrivilegeEscalation: false
runAsUser: 0
startupProbe:
failureThreshold: 40
httpGet:
path: /api/health
port: 3460
scheme: HTTP
initialDelaySeconds: 10
periodSeconds: 10
successThreshold: 1
timeoutSeconds: 1
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
dnsPolicy: ClusterFirst
restartPolicy: Always
schedulerName: default-scheduler
securityContext:
fsGroup: 1001
serviceAccount: minimal-delegate
serviceAccountName: minimal-delegate
terminationGracePeriodSeconds: 600
```
***
### Harness Terraform Provider
Go to [create Harness Delegate (DDCR)](https://registry.terraform.io/providers/harness/harness/latest/docs/resources/chaos_infrastructure) to create a Harness Delegate using [Harness Terraform provider](/harness-ai/use-harness-platform/automation/terraform-provider/harness-terraform-provider-overview.md).
### Next Steps
* [Deploy Delegate on a Cluster](/resilience-testing/chaos-engineering/use-chaos-engineering/infrastructure/types/ddcr/dedicated-delegate.md)
* [Use mTLS with Delegate](/resilience-testing/chaos-engineering/use-chaos-engineering/infrastructure/types/ddcr/mtls-support.md)
{% @harness-feedback/feedback %}