## Introduce new chaosexperiments and workflowtemplates to nc-helm charts

Helm Chart

A helm chart is a collection of kubernetes yaml manifests grouped into a single package to simplify the deployment of a containarized application.

Use tree command inside chart folder to see the chart structure

├── Chart.yaml
├── charts
├── templates
│   ├── NOTES.txt
│   ├── _helpers.tpl
│   ├── deployment.yaml
│   ├── ingress.yaml
│   ├── service.yaml
│   └── tests
│   └── test-connection.yaml
└── values.yaml

Helm chart contains,

**charts**: Used for adding dependent charts.
**templates**: Contains configuration files
**templates/tests**: Used for adding test-manifests to test helm chart
**values.yaml**: Used to define default values for configuration files

#### chaosexperiment helm chart

Chaos experiment contain the actual chaos details. These experiments are installed on target kubernetes cluster as Kubernetes CRs.

$ nc-helm/chaosexperiments# tree
.
├── Chart.yaml
├── templates
│   ├── argo-workflow-rbac.yaml
│   ├── experiment-pod-cpu-hog.yaml
│   ├── experiment-pod-delete.yaml
│   ├── experiment-pod-memory-hog.yaml
│   ├── experiment-pod-network-latency.yaml
│   ├── experiment-pod-network-loss.yaml
│   ├── litmus-chaos-rbac.yaml
│   ├── test-chaosexperiments-serviceaccount.yaml
│   └── tests
│   ├── test-chaosexperiments-pod-cpu-hog.yaml
│   ├── test-chaosexperiments-pod-delete.yaml
│   ├── test-chaosexperiments-pod-memory-hog.yaml
│   ├── test-chaosexperiments-pod-network-latency.yaml
│   └── test-chaosexperiments-pod-network-loss.yaml
└── values.yaml

templates folder contains the list of experiments available in the chart and test folder contains test-experiments used to test helm chart.

To install the experiment, edit the values.yaml file and set the experiments to true and run helm install or override the values.yaml using --set command from helm command line.

# Using values.yaml
$ cat values.yaml
experiments:
pod_delete:
enabled: false ---> set to true
network_latency:
enabled: false
network_loss:
enabled: false
pod_memory_hog:
enabled: false
pod_cpu_hog:
enabled: false

# Using --set to the helm install command
$ cd chaosexperiments
$ helm install . --name=chaosexperiments --namespace=aqua --set experiments.pod_delete.enabled=true,experiments.network_latency.enabled=true,experiments.network_loss.enabled=true,experiments.pod_cpu_hog.enabled=true,experiments.pod_memory_hog.enabled=true

By default, All chaosexperiments are set to false in values.yaml under experiments section.

Example chaosexperiment manifest template:

{{ if .Values.experiments.<experiment-name>.enabled }} -----> Enable the experiment in values.yaml
apiVersion: litmuschaos.io/v1alpha1
description:
message: |
<Description of experiment>
kind: ChaosExperiment
metadata:
name: <experiment-name>
namespace: {{ .Values.namespace }} -----> Namespace of the experiment
labels:
name: <label>
app.kubernetes.io/part-of: litmus
app.kubernetes.io/component: chaosexperiment
app.kubernetes.io/version: 1.10.0
version: 0.1.19
spec:
definition:
scope: Namespaced
image: {{ .Values.images.tags.go_runner }} -----> Docker image used to perform chaos
imagePullPolicy: {{ .Values.images.pull_policy }} -----> Image download policy
args:
- -c
- ./experiments -name <experiment-name>
command:
- /bin/bash
env:
- name: <key1>
value: <value1>
- name: <key2>
value: <value2>
labels:
name: <label>
app.kubernetes.io/part-of: litmus
app.kubernetes.io/component: experiment-job
app.kubernetes.io/version: 1.10.0
{{ end }}

To add a new experiment to chaosexperiment helm chart,

1. Create a new experiment file in templates folder with name experiment-<experiment-name>.yaml
chaosexperiment menifests can be downloaded from https://litmuschaos.io/
2. Create a test experiment with test-chaosexperiments-<experiment-name>.yaml
3. Add new experiment to values.yaml in experiments section

### litmus-workflowtemplates helm chart

Litmus use argo workflow templates. This allow developers to reuse them by referencing them from your Workflows.

litmus-argo-workflow-templates chart directory structure

├── Chart.yaml
├── templates
│   ├── chaos_manager_service_account.yaml
│   ├── pod-cpu-hog-template.yaml
│   ├── pod-delete-template.yaml
│   ├── pod-memory-hog-template.yaml
│   ├── pod-network-latency-template.yaml
│   ├── pod-network-loss-template.yaml
│   ├── test-workflowtemplates-serviceaccount.yaml
│   └── tests
│   ├── test-workflowtemplate-pod-cpu-hog.yaml
│   ├── test-workflowtemplate-pod-delete.yaml
│   ├── test-workflowtemplate-pod-memory-hog.yaml
│   ├── test-workflowtemplate-pod-network-latency.yaml
│   └── test-workflowtemplate-pod-network-loss.yaml
└── values.yaml

To install the argo-workflowtemplates, edit the values.yaml file and set the template to true and run helm install or override the values.yaml using --set command from helm command line.

# Using values.yaml
$ cat values.yaml
workflowTemplates:
pod_delete:
enabled: false ---> set to true
network_latency:
enabled: false
network_loss:
enabled: false
pod_memory_hog:
enabled: false
pod_cpu_hog:
enabled: false

# Using --set to the helm install command
$ cd litmus-argo-workflow-templates
$ helm install . --name=workflowtemplates --namespace=aqua --set workflowTemplates.pod_delete.enabled=true,workflowTemplates.network_latency.enabled=true,workflowTemplates.network_loss.enabled=true,workflowTemplates.pod_cpu_hog.enabled=true,workflowTemplates.pod_memory_hog.enabled=true

To add a new litmus-argo-workflowtemplate, create a manifest <experiment-name>-template.yaml

Example workflowtemplate:

{{ if .Values.workflowTemplates.<experiment-name>.enabled }} -----> Add the template in values.yaml
apiVersion: argoproj.io/v1alpha1
kind: WorkflowTemplate
metadata:
name: <experiment-name>
spec:
arguments:
parameters:
- name: target_namespace -----> Define input parameter here
- name: pod_label
templates:
- name: <experiment-name>
inputs:
parameters:
- name: targt_namespace -----> Define input parametes here
- name: pod_label
resource: # indicates that this is a resource template
action: create # can be any kubectl action (e.g. create, delete, apply, patch)
successCondition: status.engineStatus == completed
failureCondition: status.engineStatus == failed
manifest: | #put your kubernetes spec here
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
generateName: <experiment-name>- -----> It generates a randon name for chaosengine resource
namespace: {{ .Values.chaos_namespace }} -----> namespace for the resource
ownerReferences:
- apiVersion: argoproj.io/v1alpha1
blockOwnerDeletion: true
controller: true
kind: Workflow
name: "{{`{{workflow.name}}`}}"
uid: "{{`{{workflow.uid}}`}}"
spec:
appinfo:
appns: '{{`{{inputs.parameters.target_namespace}}`}}'
applabel: '{{`{{inputs.parameters.pod_label}}`}}'
# It can be true/false
annotationCheck: 'false'
# It can be active/stop
engineState: 'active'
#ex. values: ns1:name=percona,ns2:run=nginx
auxiliaryAppInfo: ''
chaosServiceAccount: {{ .Values.chaos_service_account }}
monitoring: false
# It can be delete/retain
jobCleanUpPolicy: '{{`{{inputs.parameters.cleanup_policy}}`}}'
components:
runner:
{{- with .Values.nodeSelector }}
nodeSelector:
{{- toYaml . | nindent 18 }}
{{- end }}
experiments:
- name: pod-delete
spec:
components:
{{- with .Values.nodeSelector }}
nodeSelector:
{{- toYaml . | nindent 22 }}
{{- end }}
env:
# set chaos duration (in sec) as desired
- name: TOTAL_CHAOS_DURATION
value: '{{`{{inputs.parameters.duration}}`}}'

# set chaos interval (in sec) as desired
- name: CHAOS_INTERVAL
value: '{{`{{inputs.parameters.chaos_interval}}`}}'

# pod failures without '--force' & default terminationGracePeriodSeconds
- name: FORCE
value: '{{`{{inputs.parameters.force}}`}}'
{{ end }}

The workflowtemplate can be tested by creating a simple workflow and add templatereference to it

$ cat workflow.yaml
apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
generateName: pod-network-latency-
spec:
serviceAccountName: litmus-workflow
entrypoint: pod-network-latency
templates:
- name: pod-network-latency
steps:
- - name: pod-network-latency
templateRef: -----> template reference
name: pod-network-latency -----> Template name
template: pod-network-latency -----> Tempalte name
arguments:
parameters:
- name: target_namespace
value: "default"
- name: pod_label
value: "name=networkchaos"

Make sure argo and litmus pods running
Install the chaosexperiment and run
$ kubectl apply -f workflow.yaml -n aqua

Verify the workflow
$ kubectl get wf -n aqua
NAME STATUS AGE
pod-network-latency-br26d Succeeded 4m50s

Running helm test for chaosexperiment

$ helm ls
NAME REVISION UPDATED STATUS CHART APP VERSION NAMESPACE
chaosexperiments 1 Fri May 28 04:36:52 2021 DEPLOYED chaosexperiments-0.1.0 1.0 aqua
litmus 1 Fri May 28 04:16:20 2021 DEPLOYED litmus-1.11.0 1.11.0 aqua
workflowtemplates 1 Fri May 28 04:37:11 2021 DEPLOYED litmus-argo-workflow-templates-0.1.0 1.0 aqua

$ helm test chaosexperiments
RUNNING: test-pod-network-loss-chaosexperiment
PASSED: test-pod-network-loss-chaosexperiment
RUNNING: test-pod-network-latency-chaosexperiment
PASSED: test-pod-network-latency-chaosexperiment
RUNNING: test-pod-delete-chaosexperiment
PASSED: test-pod-delete-chaosexperiment

Running helm test for litmus-argo-workflow-templates

$ helm test workflowtemplates
RUNNING: test-pod-network-loss-workflowtemplate
PASSED: test-pod-network-loss-workflowtemplate
RUNNING: test-pod-network-latency-workflowtemplate
PASSED: test-pod-network-latency-workflowtemplate
RUNNING: test-pod-delete-workflowtemplate
PASSED: test-pod-delete-workflowtemplate