Usage

Krkn Operator documentation v1.1.0; captured October 13, 2026 from website commit 3b059252814d83dc2ff56caa6031603229ed9cdf.

• 1: Jobs
• 2: Cluster Terminal
• 3: Run Scenarios
• 4: Chaos Studio
• 5: File Management
• 6: Job Categories & Resiliency Score Report

1 - Jobs

Jobs View Cancel

The Jobs page is the platform home screen. It lists scenario executions available to your group, with status updates and access to logs and results.

Run Types

Single Run

A single scenario execution on one or more target clusters. For each run you can inspect:

• Execution outcome (success, failure)
• Console logs of the krkn pod that executed the scenario

Replay & Edit

Single runs can be replayed to re-execute the same scenario with the same configuration. You can also edit the details before replaying to modify:

• Target cluster(s)
• Scenario parameters
• Configuration values
• Run Name

This allows you to quickly iterate on chaos experiments or run the same test across different environments.

Graph Run (Chaos Studio)

A workflow execution created through Chaos Studio. The execution is displayed as a graph where each node represents an individual krkn scenario. For each node you can inspect the outcome and logs independently.

Graph runs can include a Resiliency Score — a calculated metric based on PromQL queries executed during the workflow. This score measures how well the target application withstood the chaos experiment.

Visibility & Permissions

• Only jobs from your own group are displayed
• A cluster-wide tip shows the names of scenarios currently running across all groups (no details)
• Cancelling or removing jobs requires Cancel permission

2 - Cluster Terminal

Cluster Terminal Run

A built-in terminal for inspecting target clusters with read-only oc and kubectl commands before, during, or after chaos experiments.

Details

• Read-only: only read commands are available — no modifications to the cluster are possible
• Supported commands: oc and kubectl
• Target selection: only clusters accessible to your group are available
• Requires Run permission

3 - Run Scenarios

Run Scenarios Run

Use the guided wizard to execute a chaos scenario on one or more target clusters in parallel.

Execution Flow

1. Select Target Clusters

Choose one or more clusters to run the scenario on. All selected clusters execute the scenario simultaneously.

2. Select Registry

Choose the container registry for scenario images:

For a disconnected target, follow Mirror Signed Images before selecting a private registry. Scenario signatures and any helper workload images must be available in the mirror.

3. Select Scenario

Browse and select a chaos scenario from the chosen registry.

4. Configure Parameters

Each scenario defines its own parameter set, divided into three categories:

When a scenario starts a helper workload, set its IMAGE parameter to the mirrored workload-* image URL. Selecting a private scenario registry does not change the workload image automatically. See Mirror Signed Images for an example.

Elasticsearch in Global Parameters

When configuring Global parameters, you can select a saved Elasticsearch endpoint to index scenario metrics and telemetry:

1. Expand the Global parameters section
2. Locate the Elasticsearch dropdown
3. Select from available saved configurations (configured by your admin)

Selecting an Elasticsearch configuration automatically applies connection details (URL, index, credentials) without requiring manual input.

Load Cloud Credential

When a scenario declares cloud-related fields (CLOUD_TYPE, AWS_*, AZURE_*, and similar), a Load Cloud Credential section appears:

1. Open Load Cloud Credential
2. Select a saved credential from the dropdown (configured by your admin)
3. Matching cloud fields become read-only and show masked placeholders (••••••••)
4. Fields belonging to other cloud providers are hidden from the form
5. Continue configuring non-cloud parameters as usual

The dropdown only lists credentials your group can access.

Selecting a credential stores only the credential name on the run (cloudCredentialRef). The console and API strip plaintext cloud environment variables from the payload so they cannot land in the Custom Resource. The operator injects secret values into the scenario pod via SecretKeyRef.

5. Mount Files

Optionally attach configuration files previously uploaded through File Management. Files are available via a select dropdown.

6. Preview

Review a summary of all configured parameters before execution.

7. Run

Launch the scenario. You are redirected to the Jobs page where you can monitor the execution in real time.

4 - Chaos Studio

Chaos Studio Run

A visual drag-and-drop editor for designing complex chaos workflows. Build graphs of scenarios that execute in series or parallel, and measure application resilience with the Resiliency Score.

Creating and Saving Workflows

Creating a New Workflow

1. Navigate to Chaos Studio from the main navigation
2. Click the “New Workflow” button to start a blank canvas
3. Give your workflow a meaningful name that describes its purpose

Adding Nodes

1. Click the “Add Node” button or drag from the node palette
2. Position nodes on the canvas by dragging them
3. Configure each node by clicking on it and selecting:
   • Registry: Choose your scenario source
   • Scenario: Select the chaos scenario to run
   • Parameters: Set mandatory and optional parameters
   • Cloud credential: Optionally load a saved cloud credential for cloud-dependent scenarios
   • Files: Mount any required configuration files

Connecting Nodes

• Click and drag from a node’s output port to another node’s input port to create dependencies
• Serial execution: connect nodes in a chain (A → B → C)
• Parallel execution: connect one node to multiple targets (A → B and A → C)

Saving Your Workflow

1. Click the “Save” button in the top toolbar
2. The workflow is automatically saved to your workspace
3. Saved workflows appear in the Chaos Studio workflow list
4. You can edit saved workflows by selecting them from the list

Visual Editor

• Nodes represent individual krkn scenarios
• Arrows define execution order — connect nodes to create dependencies
• Serial execution: chain nodes sequentially with single arrows
• Parallel execution: branch from one node to multiple targets
• Drag and drop to arrange the workflow visually

Node Configuration

Each node is configured exactly like a single scenario: select a registry, choose a scenario, set mandatory/optional/global parameters, optionally load a cloud credential, and mount files.

Global Parameters in Workflows

When configuring global parameters for workflow nodes:

• Elasticsearch: Select from saved admin-configured endpoints to index metrics across the entire workflow
• Prometheus: Configure Prometheus integration for metric collection
• Cerberus: Enable cluster health monitoring during scenario execution

Cloud Credentials in Workflows

Cloud-dependent nodes can use saved credentials configured under Cloud Credentials Management.

When configuring a node:

1. Open Load Cloud Credential
2. Select a credential available to your group
3. Cloud fields for that provider become masked; other providers’ fields are hidden

Each node can select its own credential (for example one AWS credential on a zone-outage node and a BMC credential on a power-outage node).

Clone: duplicate a configured node to repeat the same scenario multiple times within the workflow.

Validation

The editor validates the workflow before execution:

• Configured nodes: green arrows and connections — ready to execute
• Unconfigured nodes: red arrows and red node border — must be configured before the workflow can run

Only fully configured nodes can be connected and executed.

Resiliency Score

Enable the Resiliency Score on a workflow to measure application resilience:

1. Enable the resiliency score option on the workflow
2. Mount a PromQL file — a special file containing PromQL queries uploaded via File Management
3. Queries execute during the workflow and their results contribute to the final score
4. View the score in the Jobs page alongside the graph run results

The PromQL file defines which metrics to evaluate (latency, error rate, throughput, etc.) and their expected thresholds. The score reflects how closely the application maintained its expected behavior under chaos.

5 - File Management

File Management

Create and manage text-based configuration files for scenario executions and Chaos Studio workflows.

Adding Files

Create a file by pasting its contents into the text input box. The console does not upload files from your device.

Where Files Are Used

6 - Job Categories & Resiliency Score Report

Job Categories & Resiliency Score Report

Use categories to organize scenario runs, then compare their resiliency scores over time in the report.

Create a Category

1. Open Categories from the left navigation menu.
2. Select Create Category.
3. Enter a name and choose a color in the color picker.
4. Set the category visibility and create it.

Categories help you group related runs for reporting. Choose a color to identify the category in the console.

Assign a Category to Runs

1. Open Jobs and locate a scenario run.
2. Open the run’s Actions menu and select Category.
3. Select one or more categories to assign them to the run.
4. Repeat for other runs you want to compare.

The category is saved on the run and can be used as a filter in the report.

For example, assign score-category to pod-scenarios-99629e65 and pod-scenarios-aedd3317 to compare their resiliency scores together.

View the Resiliency Score Report

1. Open Resiliency History from the left navigation menu.
2. Select one or more categories and target clusters.
3. Select Apply filters to display the report.

The report plots resiliency scores over time. In Separate configuration groups mode, runs with the same settings appear in the same chart; runs with different settings appear in separate charts. This makes it easier to compare like-for-like runs while keeping different configurations distinct.

Combine Configurations by Category

Select Combine configurations by category to put runs from the selected category into a single chart, including runs with different settings. This helps compare scores across heterogeneous configurations.

When the report includes more than one cluster, each cluster is shown as a separate line so you can compare results across clusters. The report can also be exported as a PDF.