> 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/disaster-recovery-testing/pipeline-stage-reference.md).
# Pipeline Stage Reference
Complete field reference for Disaster Recovery (DR) test pipelines. For a step-by-step walkthrough, go to [Get Started with DR Testing](/resilience-testing/disaster-recovery-testing/get-started.md).
{% hint style="info" %}
**FEATURE FLAG**
DR Testing is behind a feature flag (`CHAOS_DR_TESTING_ENABLED`). Contact your Harness sales representative to enable it for your account.
{% endhint %}
***
### DR Tests list
Go to **Resilience Testing → DR Tests**. The list shows pipelines that contain at least one Disaster Recovery stage.
| Column | Description |
| --------------------- | ---------------------------------------------------------- |
| **Pipeline name** | Display name and auto-generated identifier. |
| **DR test stages** | Names of the `DRTest` stages in the pipeline. |
| **Recent executions** | Sparkline of recent run statuses. |
| **Last execution** | Status and timestamp of the most recent run, or **Never**. |
| **Last modified** | Timestamp and user who last saved the pipeline. |
Filterable by **Search** and **Tag(s)**. Default sort: **Last Modified (New -> Old)**.
***
### Create new DR test dialog
Shown when you select **+ New DR Test**.
| Field | Required | Description |
| --------------- | -------------- | --------------------------------------------------------------------------------------------------------- |
| **Name** | Yes | Display name for the Disaster Recovery stage. Also used to derive the pipeline name (` Pipeline`). |
| **Id** | Auto-generated | Unique identifier derived from the name. Editable via the pencil icon. Must be unique within the project. |
| **Description** | No | Free-text description of the disaster scenario. |
| **Tags** | No | Key or `key:value` labels for organizing and filtering. |
| **Objective** | No | Goal or success criteria for this DR test. |
Select **Continue in Pipeline Studio** to create the pipeline and open it. Select **Cancel** to discard the dialog without creating anything.
Harness creates:
* A pipeline with identifier derived from the name and tag `module: drtest`.
* A stage of type `DRTest` whose name and identifier match the dialog values.
* An empty Execution canvas, so you add Chaos Probe, Chaos Fault, Chaos Action, and other steps yourself.
***
### Pipeline Studio: Overview tab
The stage overview is pre-populated from the create dialog. All fields remain editable.
#### Stage overview fields
| Field | Description |
| --------------- | --------------------------------------------------- |
| **Name** | Stage display name. |
| **Id** | Stage identifier. Auto-derived from name, editable. |
| **Description** | Optional stage description. |
| **Tags** | Optional labels. |
| **Objective** | Optional objective for this disaster recovery test. |
#### Advanced section (within Overview)
| Field | Description |
| ------------------- | ----------------------------------------------------------------------------------------------------------------------------- |
| **Timeout** | Maximum allowed runtime for the stage. Accepts duration format: `w`, `d`, `h`, `m`, `s`. Example: `2h 30m`. |
| **Stage Variables** | Key-value variables scoped to this stage. Select **+ New Variable** to add. Reference them in steps with Harness expressions. |
***
### Pipeline Studio: Environment tab
#### Configuration section
| Field | Description |
| ----------------------- | -------------------------------------------------------------------------------------------- |
| **Specify Environment** | Select the Harness environment for the stage, or select **+ New Environment** to create one. |
Chaos infrastructure is **not** selected on this tab. Each Chaos Fault, Chaos Probe, or Chaos Action step selects its own infrastructure.
#### Failure Strategy section
Defines how the pipeline responds when the stage encounters an error. You can add multiple strategies for different error types.
**On failure of type**
Select one or more failure types, or check **All Errors**:
| Failure type | Description |
| ------------------------------------ | ------------------------------------------------------- |
| Authentication Errors | Invalid or expired credentials. |
| Connectivity Errors | Network or connectivity issues reaching a target. |
| Timeout Errors | Operations that exceed their configured timeout. |
| Authorization Errors | Insufficient permissions. |
| Verification Failures | Step or probe verification conditions not met. |
| Delegate Provisioning Errors | Failures provisioning or connecting a Harness Delegate. |
| Unknown Errors | Errors that do not match any other category. |
| Policy Evaluation Failures | Governance policy checks that fail. |
| Execution-time Inputs Timeout Errors | Runtime input prompts that were not fulfilled in time. |
| Approval Rejection | An approval step was explicitly rejected. |
| Delegate Restart | The Delegate restarted during execution. |
| User Marked Failure | A step was manually marked as failed. |
**Perform Action**
| Action | Description |
| --------------------- | ------------------------------------------------------------------------ |
| **Rollback Pipeline** | Roll back all stages in the pipeline that support rollback. |
| **Retry Step** | Retry the failed step a configured number of times. |
| **Abort** | Stop pipeline execution immediately. |
| **Mark As Failure** | Mark the stage as failed. |
| **Rollback Stage** | Roll back only this stage, using the Rollback path on the Execution tab. |
***
### Pipeline Studio: Execution tab
The Execution tab has two canvases, toggled with **Execution | Rollback**:
* **Execution:** The forward recovery workflow.
* **Rollback:** Compensating steps for when a failure strategy selects **Rollback Stage** or **Rollback Pipeline**. Harness does not enter this path automatically on probe failure unless your failure strategy invokes rollback.
Select **Add Step**, then choose **Add Step**, **Add Step Group**, **Use template**, or **Create with AI**.
In the **Step Library**, open **Resilience Testing** to add Chaos Fault, Chaos Probe, Chaos Action, or Chaos Experiment. Open **Approval** to add Harness Approval or other approval steps. An empty Disaster Recovery stage can be saved; approval, fault, and probe steps are available but not required by schema validation before save.
#### Chaos Probe
Validates a condition against your system for a configured duration.
| Field | Required | Description |
| ------------------------------- | ---------------- | ------------------------------------------------------------------------------------------------------------------------- |
| **Name** | Yes | Display name for the step. |
| **Id** | Auto-generated | Step identifier. |
| **Chaos Probe** | Yes | The probe identity to run, such as an HTTP or Kubernetes health probe from your project. |
| **Select Chaos Infrastructure** | Yes | Target infrastructure in the form `/`. |
| **Duration** | Yes | How long the probe runs (for example `1m`, `10m`). Placeholder text shows examples such as `10s`, `1m`, `1h`. |
| **Runtime Inputs** | Depends on probe | Name and value pairs the selected probe exposes. Required inputs are marked. Pin a field to supply the value at run time. |
YAML shape:
```yaml
- step:
type: ChaosProbe
name: ChaosProbe_1
identifier: ChaosProbe_1
spec:
identity: ad-http-test-1
duration: 10m
infraReference: env91x/infra91xadvanced
tasks:
- identifier: ad-http-test-1
values:
- name: VARIABLES_URL
value: https://example.com
```
#### Chaos Fault
Injects a failure into the target system.
| Field | Required | Description |
| ------------------------------- | ---------------- | ------------------------------------------------------------------------------------------------------------------- |
| **Name** | Yes | Display name for the step. |
| **Id** | Auto-generated | Step identifier. |
| **Chaos Fault** | Yes | Fault identity to inject. This can be a project fault or a template-backed identity such as `custom.`. |
| **Select Chaos Infrastructure** | Yes | Target infrastructure in the form `/`. |
| **Runtime Inputs** | Depends on fault | Inputs the selected fault exposes. Pin fields you want prompted at run time. |
YAML shape:
```yaml
- step:
type: ChaosFault
name: ChaosFault_1
identifier: ChaosFault_1
spec:
identity: custom.template-scope-test-eps
infraReference: env91x/infra91xadvanced
```
#### Chaos Action
Runs a predefined chaos action from Resilience Testing.
| Field | Required | Description |
| ------------------------------- | ----------------- | ------------------------------------------------------------------------------------------------- |
| **Name** | Yes | Display name for the step. |
| **Id** | Auto-generated | Step identifier. |
| **Chaos Action** | Yes | The action identity to run. |
| **Select Chaos Infrastructure** | Yes | Target infrastructure in the form `/`. |
| **Duration** | Yes | How long the action runs. |
| **Runtime Inputs** | Depends on action | Inputs the selected action exposes, such as `VARIABLES_DURATION`. Pin fields for runtime prompts. |
YAML shape:
```yaml
- step:
type: ChaosAction
name: ChaosAction_1
identifier: ChaosAction_1
spec:
identity: k8s-delay-variable
duration: 10s
infraReference: env91x/infra91xadvanced
tasks:
- identifier: k8s-delay-variable
values:
- name: VARIABLES_DURATION
value: 10s
```
You can also add standard Harness steps (shell script, HTTP, Harness Approval, notification) in the same stage.
#### Use templates and Save as Template
* **Use template:** From the Add Step menu, opens **Step and Step Group Templates**. Select a template to insert it into the Execution (or Rollback) canvas. Template-backed faults often appear as an `identity` such as `custom.` in YAML.
* **Save as Template:** On a Chaos Probe, Chaos Fault, or Chaos Action drawer, saves the current step configuration as a reusable template.
#### Variables
Open **Variables** in Pipeline Studio to manage:
* **Pipeline** variables (name, identifier, tags, custom variables).
* **Stage** variables for the Disaster Recovery stage (**Add Variable** under Stage Variables).
Reference them from step inputs with Harness expressions such as `<+stage.variables.myVar>`.
***
### Pipeline Studio: Advanced tab
#### Delegate Selector
| Field | Description |
| ---------------------------- | -------------------------------------------------------------------------------------------------------------------------- |
| **Define Delegate Selector** | Optional. Add selector tags to route this stage to a specific Harness Delegate. Leave empty to use any available Delegate. |
#### Conditional Execution
| Option | Description |
| ------------------------------------------------------ | ------------------------------------------------------------- |
| If the pipeline executes successfully up to this point | Default. Stage runs only if preceding stages passed. |
| If the previous pipeline or stage fails | Stage runs only if the immediately preceding stage failed. |
| Always | Stage always runs. |
| JEXL Condition | Stage runs only if the custom expression evaluates to `true`. |
#### Looping Strategy
| Strategy | Description |
| --------------- | --------------------------------------------------------------------------- |
| **Matrix** | Run the stage for each combination of values across variable axes. |
| **Repeat** | Run the stage a fixed number of times, optionally with a concurrency limit. |
| **Parallelism** | Split stage iterations to run in parallel up to a configured maximum. |
#### Failure Strategy (Advanced tab)
Same failure type options as the Environment tab. Additional actions available here:
| Action | Description |
| ------------------------------- | ------------------------------------------------------------- |
| **Manual Intervention** | Pause and wait for a user to choose the next action. |
| **Ignore Failure** | Treat the failure as a warning and continue. |
| **Retry Step** | Retry the failed step. |
| **Mark As Success** | Override the failure and mark the step or stage as succeeded. |
| **Abort** | Stop execution immediately. |
| **Proceed with Default Values** | Continue using default values for unresolved inputs. |
| **Mark As Failure** | Mark the step or stage as failed. |
| **Rollback Pipeline** | Roll back all stages that support rollback. |
***
### Resilience Tests tab on execution
On a pipeline execution, open **Resilience Tests**. The tab lists Resilience Testing steps from the selected stage and reports **Probe Results** for probe outcomes.
* Use it to review probe pass or fail results after a DR run.
* For Chaos Fault and Chaos Action detail, use step **Details**, **Step Logs**, and **Console View**. The tab is not a fault catalog.
***
### Related
* [Get Started with DR Testing](/resilience-testing/disaster-recovery-testing/get-started.md): Create and run your first DR test.
* [Concepts](/resilience-testing/disaster-recovery-testing/concepts.md): RTO, RPO, environments, and failure strategies.