> 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-testing/infrastructure.md).

# Infrastructure

### What you will learn <a href="#what-you-will-learn" id="what-you-will-learn"></a>

* What a chaos infrastructure is and why it is needed.
* The three supported types and how they map to the **Project Settings → Resilience Testing Infrastructures** UI.
* Where to go next to install, secure, and upgrade infrastructure.

***

### Where infrastructures live in the UI <a href="#where-infrastructures-live-in-the-ui" id="where-infrastructures-live-in-the-ui"></a>

Chaos infrastructures are managed per project under **Resilience Testing → Project Settings → Resilience Testing Infrastructures**. The page has one tab per supported type:

* **Kubernetes (Harness Infrastructure)**
* **Linux**
* **Windows**

Each infrastructure is scoped to a Harness **environment** (project, organization, or account-scoped). One environment can have multiple chaos infrastructures attached.

{% hint style="info" %}
**DEPRECATED: KUBERNETES (DEDICATED CHAOS INFRASTRUCTURE)**

You may also see a **Kubernetes (Dedicated Chaos Infrastructure)** tab. This is the legacy chaos-agent-based topology and is deprecated. New infrastructures cannot be created on it, and existing ones should be migrated to **Kubernetes (Harness Infrastructure)** (the DDCR flow) when possible.
{% endhint %}

{% hint style="info" %}
**DELETING ENVIRONMENTS**

To delete an environment, first detach or delete every chaos infrastructure attached to it. The environment delete will fail while a referenced infrastructure exists.
{% endhint %}

***

### Supported infrastructure types <a href="#supported-infrastructure-types" id="supported-infrastructure-types"></a>

Harness Chaos Testing supports three chaos infrastructure types. **Kubernetes** is recommended for all new setups; the others cover non-Kubernetes targets.

| Type           | Target                                               | Doc                                                                                                |
| -------------- | ---------------------------------------------------- | -------------------------------------------------------------------------------------------------- |
| **Kubernetes** | Kubernetes clusters, via the Harness Delegate (DDCR) | [Set up Kubernetes infrastructure](/resilience-testing/chaos-testing/infrastructure/kubernetes.md) |
| **Linux**      | Linux VMs and Cloud Foundry                          | [Linux infrastructure](/resilience-testing/chaos-testing/infrastructure/linux.md)                  |
| **Windows**    | Windows VMs                                          | [Windows infrastructure](/resilience-testing/chaos-testing/infrastructure/windows.md)              |

Chaos experiments against AWS, GCP, Azure, VMware, and bare-metal targets run **on top of** a Kubernetes chaos infrastructure (the runner pods execute there even when the fault hits a non-Kubernetes resource). Cloud Foundry experiments run on Linux infrastructure.

***

### Next steps <a href="#next-steps" id="next-steps"></a>

* [Set up Kubernetes infrastructure](/resilience-testing/chaos-testing/infrastructure/kubernetes.md): install the Delegate-driven runner and connect a target cluster.
* [Cluster permissions](/resilience-testing/chaos-testing/infrastructure/kubernetes/permissions.md): review the Kubernetes RBAC the chaos service account needs.
* [Network configuration](/resilience-testing/chaos-testing/infrastructure/kubernetes/network-config.md): mTLS and Harness Network Proxy (HNP) settings for Kubernetes chaos infrastructure.
* [Upgrade Linux infrastructure](/resilience-testing/chaos-testing/infrastructure/linux/upgrade.md): upgrade an existing Linux install across releases.
* [Upgrade Windows infrastructure](/resilience-testing/chaos-testing/infrastructure/windows/upgrade.md): upgrade an existing Windows install across releases.

Before creating an infrastructure, [create an environment](/resilience-testing/chaos-engineering/chaos-experiments/create-experiments.md#create-environment). A chaos infrastructure always lives inside an environment.