NOTE: These steps assumes you already have an account in Harness, if not, please create a new project or ask your administrator to add you to an existing project in Harness.

Please follow the below guidelines to run your first Chaos experiment Pod Delete:

Sequence in which the APIs should be executed to run a sample chaos experiment such as Pod Delete (all the APIs to do so are present in folder "Run Pod Delete Chaos Experiment" under Harness CE Postman Collection).

Listed below are the required inputs and how to access (or generate) them. We will refer these variables in the API requests explained in this tutorial -

url

This is the Harness CE server URL serving the Chaos API requests. You can find this URL in the Chaos API doc(top right). For instance, https://app.harness.io/gateway/chaos/manager/api/query is the Harness CE server URL for production environment. If you are using Chaos postman collection, this value will be pre-filled in 2 postman variables like below:

• base_url - https://app.harness.io
• chaos_url - $base_url/gateway/chaos/manager/api/query

account_id

This is your Harness account Id. Please use the account where you would like to run Chaos experiments.

You can retrieve your account Id by following the below steps:

1. Go to ACCOUNT SETTINGS in Harness.
2. Click on Overview.
3. Copy the value of Account Id.

project_id

This is your Harness project Id. Please use the project where you would like to run Chaos experiments.

You can retrieve your project Id by following the below steps:

1. Go to Projects in Harness.
2. Select the project where you would like to run the Chaos experiments or create a new project.
3. Click on Overview.
4. Copy the value of Id.
5. Please make sure to copy the value of Organization as well, this will be the value of next required variable org_id.

org_id

This is your Harness organization Id. Please use the organization where you would like to run Chaos experiments.

You can retrieve your organization Id by following the below steps: See above step to retrieve project_id - you would have copied the value of org_id as well in that step, if not please follow the same steps again to retrieve organization Id

API-KEY-TOKEN

You can use an existing Harness API key token of the same Harness account selected in above steps or you can create a new one by following the below steps:

1. Click on MY PROFILE in Harness.
2. Go to My API Keys section and click on + API Key button to create a new API Key.
3. Enter a Name for the API Key(optional Description/Tags) and click on Save.
4. Under the newly created API Key, click on + Token button to generate a new token for this API Key.
5. Enter the Name and select the Expiration(time for which this token will be valid) and click on Generate Token.
6. Please make sure to copy the value of the token generated and store somewhere safe, you may not be able to copy this later.

Variables In Harness CE Postman Collection

In order to work with Harness CE Postman Collection, please fork it under your own Postman account and update the above-mentioned variables in order to work with Harness CE Postman Collection.

Authorization Header For API Request

An authorization header with API-KEY-TOKEN value retrieved above is required to work with Chaos APIs i.e., x-api-key: $API-KEY-TOKEN

APIs to be executed in given sequence:

RegisterInfra

Now, we can create/register a chaos infrastructure. Here, we will add a Kubernetes infrastructure so that we can inject Kubernetes resource faults.

Add a name to your chaos infrastructure and optionally a description and tags. After that, choose the mode of installation for Harness delegate. As a quick primer, Harness delegate is a remote agent for accessing your Kubernetes cluster resources and injecting faults into them as part of a chaos experiment. The Cluster Wide installation mode allows you to target resources across all the namespaces in your cluster while Namespace Mode installation restricts chaos injection to only the namespace in which the delegate will be installed. In this tutorial, the delegate will be installed in the hce namespace with Cluster Wide mode, but you can change it.

NOTE: Above values are pre-filled in Harness CE Postman Collection.

Please make sure to note down the value of field token and infraID from response, these values will be used in subsequent API calls.

Once above request is executed, retrieve the infrastructure manifest by forming URL as shown below: $chaos_server_url/file/$token.yaml where $chaos_server_url would be https://app.harness.io/gateway/chaos/manager/api/ and $token is the token value retrieved in response of registerInfra API call as shown in sample response above. An example URL to retrieve infrastructure manifest could look like https://app.harness.io/gateway/chaos/manager/api/file/token-abc.yaml , copy and save the yaml in a machine where kubectl is installed and has access to your k8s cluster in some file named infra.yaml.

Lastly, provided that you have access to your Kubernetes cluster via kubectl, deploy your chaos infrastructure by executing the below commands - kubectl apply -f infra.yaml

It will take a while for the delegate to setup in the k8s cluster. Meanwhile, if you want to check the status of your infra, please use the API GetInfraDetails as explained below.

GetInfraDetails

Optional - If you want to get the details of your newly registered infra or to check if your infra is active or not before creating a Chaos workflow, you can use GetInfraDetails API - It will return the details of a given infra based on infra ID input. Please make sure to check the value of field isActive, it should be true i.e., the infrastructure should be in active state prior to moving forward.

NOTE: Above values are pre-filled in Harness CE Postman Collection.

ListPredefinedWorkflow

Optional - Required if and only if you want to run a different chaos experiment other than pod-delete, by default, pod delete experiment manifest is present in CreateChaosWorkFlow API call) - This will list down all the available predefined workflows/experiments that can be run on your infra, please select one of them and copy the manifest from there and make necessary changes such as namespace, etc.

CreateChaosWorkFlow

Please make sure to deploy some sample target application under hce namespace with label app=cartservice prior to running this API. With our target application deployed, we can now create a chaos experiment. We will be targeting the pods of the deployed app with the Pod Delete fault. To create the chaos experiment, Add the experiment name and optionally a description and tags. Then, Add the target infrastructure id, which we created previously.

This will allow us to create our chaos experiment with a pod-delete fault configured to target the Online Boutique application.

NOTE: In order to run this API using Harness CE Postman Collection, you just need to update the value of variable infraID retrieved during registerInfra API call. The value for workflowManifest field is pre-filled in the Harness CE postman collection, user does not need to update in case they want to target an application under hce namespace with label app=cartservice mentioned in this tutorial.

RunChaosExperiment

Once a workflow is created, you can now execute the experiment by providing the workflow id from CreateChaosWorkFlow API call. NOTE: Please note down the value of field notifyID.

ListWorkflowRun

Optional - Required only if you want to check the details of the workflow currently running - details can be validated by UI as well if user prefers to do so) - It will list down all the workflow runs with the required details for those workflows, this API call can be used to validate if the workflow created by RunChaosExperiment API call is running or not, completed, etc using the field phase.

Let's go to Harness CE Postman Collection to see this in action.