Skip to main content

Pipeline Stage Reference

Last updated on

Complete field reference for all DR test pipeline stage options. For a step-by-step walkthrough, see Get Started with DR Testing.


DR Test Details Dialog

Shown when you click + New DR Test.

FieldRequiredDescription
NameYesDisplay name shown in the DR Testing list and execution history.
IdAuto-generatedUnique identifier derived from the name. Editable via the pencil icon. Must be unique within the project.
DescriptionNoFree-text description of the disaster scenario being tested.
TagsNoKey or key:value labels for organizing and filtering DR tests.
ObjectiveNoThe goal or success criteria for this DR test (e.g., "Validate regional failover completes within RTO of 15 minutes").

Click Continue in Pipeline Studio to open the pipeline stage editor. Clicking Cancel discards the dialog without creating a DR test.


Pipeline Studio: Overview Tab

The stage overview is pre-populated from the DR Test Details dialog. All fields are editable here.

Stage Overview fields

FieldDescription
NameStage display name.
IdStage identifier. Auto-derived from name, editable.
DescriptionOptional stage description.
TagsOptional labels.
ObjectiveOptional objective description for this disaster recovery test.

Advanced section (within Overview)

FieldDescription
TimeoutMaximum allowed runtime for the stage. Accepts duration format: w (weeks), d (days), h (hours), m (minutes), s (seconds). Example: 2h 30m. If the stage exceeds this duration, it is treated as a timeout failure.
Stage VariablesKey-value variables scoped to this stage. Click + New Variable to add. Variables are accessible in step configurations via expressions.

Pipeline Studio: Environment Tab

Configuration section

FieldDescription
Specify EnvironmentSelect the Harness environment where the DR test executes. Choose from existing environments in the dropdown or click + New Environment to create one.
Specify InfrastructureSelect the Chaos Infrastructure connected to the target cluster. Choose from existing infrastructure in the dropdown or click + New Infrastructure to create one. Required for Chaos Fault and Chaos Probe steps.

Failure Strategy section (Environment tab)

Defines how the pipeline responds when the stage encounters an error. You can add multiple failure strategies for different error types.

On failure of type

Select one or more failure types this strategy applies to, or check All Errors to catch all failures:

Failure TypeDescription
Authentication ErrorsFailures caused by invalid or expired credentials
Connectivity ErrorsNetwork or connectivity issues reaching a target
Timeout ErrorsOperations that exceed their configured timeout
Authorization ErrorsInsufficient permissions to perform an action
Verification FailuresStep or probe verification conditions not met
Delegate Provisioning ErrorsFailures in provisioning or connecting a Harness Delegate
Unknown ErrorsErrors that do not match any other category
Policy Evaluation FailuresHarness governance policy checks that fail
Execution-time Inputs Timeout ErrorsRuntime input prompts that were not fulfilled within the timeout
Approval RejectionAn approval step was explicitly rejected
Delegate RestartThe Delegate restarted during execution
User Marked FailureA step was manually marked as failed

Perform Action (Environment tab)

ActionDescription
Rollback PipelineRoll back all stages in the pipeline that support rollback
Retry StepRetry the failed step a configured number of times
AbortImmediately stop pipeline execution
Mark As FailureMark the stage as failed and continue or stop based on pipeline settings
Rollback StageRoll back only this stage

Pipeline Studio: Execution Tab

The Execution tab contains the step canvas where you build the DR workflow. The canvas starts with a single Add Step (+) node.

Click + to open the Step Library. Under the Disaster Recovery category, three DR-specific step types are available:

Chaos Probe

Validates the health of a Kubernetes workload.

FieldRequiredDescription
NameYesDisplay name for the step
Select Chaos InfrastructureYesThe Kubernetes Chaos Infrastructure to run the probe against
Chaos ProbeYesSelect a predefined probe (e.g., default-pod-level-probe)
DurationYesHow long the probe runs (e.g., 1m, 5m)

Runtime inputs depend on the selected probe. For default-pod-level-probe:

InputRequiredDescription
TARGET_NAMESYesName of the workload to validate (e.g., frontend)
TARGET_NAMESPACEYesKubernetes namespace of the target (e.g., boutique)
TARGET_KINDYesKubernetes resource kind (e.g., Deployment, StatefulSet)

Chaos Fault

Injects a failure into the target system.

FieldRequiredDescription
NameYesDisplay name for the step
Select Chaos InfrastructureYesThe Kubernetes Chaos Infrastructure to inject the fault on
Chaos FaultYesSelect a fault type (e.g., pod-delete, pod-network-loss, pod-cpu-hog)

Runtime inputs depend on the selected fault. For pod-delete:

InputRequiredDescription
TOTAL_CHAOS_DURATIONYesTotal duration in seconds for the fault (e.g., 60)
CHAOS_INTERVALYesInterval in seconds between fault injections (e.g., 10)
FORCENotrue for force delete, false for graceful (default: false)
RAMP_TIMENoSeconds to wait before starting the fault (default: 0)
SEQUENCENoparallel or serial pod deletion (default: parallel)
TARGET_WORKLOAD_KINDYesKubernetes resource kind (e.g., Deployment)
TARGET_WORKLOAD_NAMESPACEYesNamespace of the target workload
TARGET_WORKLOAD_NAMESYesName of the target workload
TARGET_WORKLOAD_LABELSNoLabel selectors as an alternative to names
POD_AFFECTED_PERCENTAGENoPercentage of pods to affect (default: 100)
TARGET_PODSNoSpecific pod names to target
NODE_LABELNoTarget pods running on nodes with this label

Chaos Action

Executes a predefined chaos action from your Resilience Testing module.

FieldRequiredDescription
NameYesDisplay name for the step
Select Chaos InfrastructureYesThe Chaos Infrastructure to run the action on
Chaos ActionYesSelect a predefined chaos action

Steps can also include standard Harness pipeline steps (shell scripts, HTTP calls, approval steps, notification steps, etc.) alongside DR-specific steps.


Pipeline Studio: Advanced Tab

Delegate Selector

FieldDescription
Define Delegate SelectorOptional. Add one or more selector tags to route this stage to a specific Harness Delegate. Leave empty to target any available Delegate.

Conditional Execution

Controls whether this stage runs at all, based on the state of the pipeline.

OptionDescription
If the pipeline executes successfully up to this pointStage runs only if all preceding stages passed. This is the default.
If the previous pipeline or stage failsStage runs only if the immediately preceding stage failed. Useful for rollback or alerting stages.
AlwaysStage always runs regardless of prior stage results.
JEXL ConditionDefine a custom boolean expression using JEXL. The stage runs only if the expression evaluates to true.

Looping Strategy

Optionally run the stage multiple times in a single execution. Select one strategy:

StrategyDescription
MatrixRun the stage for each combination of values across one or more variable axes.
RepeatRun the stage a fixed number of times, optionally with a concurrency limit.
ParallelismSplit 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 at the Advanced tab level:

ActionDescription
Manual InterventionPause execution and wait for a user to choose the next action
Ignore FailureTreat the failure as a warning and continue execution
Retry StepRetry the failed step
Mark As SuccessOverride the failure and mark the step or stage as succeeded
AbortStop execution immediately
Proceed with Default ValuesContinue execution using default values for any unresolved inputs
Mark As FailureMark the step or stage as failed
Rollback PipelineRoll back all stages in the pipeline

DR Tests List

ColumnDescription
DR Test NameDisplay name and auto-generated ID.
Recent ExecutionsSparkline of recent run statuses.
Last ExecutionStatus and timestamp of the most recent run.
Last ModifiedTimestamp and user who last saved the DR test.

Filterable by Infrastructures, Status, and Tag(s). Default sort: Last Modified (New to Old).