Service

After creating a project, Keptn allows creating a service or onboarding a service into Keptn. The difference between creating and onboarding is as follows:

• Create a service: This creates a new and empty service in the specified project. This option is used when not deploying a service with Keptn.

• Onboard a service: This creates a new service and uploads the configuration to deploy the service. The configuration has to be a Helm Chart.

Service name restrictions

• Service name must be a valid unix directory name (Note: for each service a directory with the corresponding name is created in the upstream git repository)
• Keptn Version <= 0.8.2: Service name in the form ${PROJECT}-${STAGE}-${SERVICE}-generated must be less than 53 characters (Note: this limitation comes from Kubernetes/Helm)
• Keptn Version >= 0.8.3: Service name must be less than 43 characters (Note: template was reduced to ${SERVICE}-generated to allow longer service names)

Create a service

• To create a service, use the keptn create service command and provide the service and project name (--project flag/):

keptn create service SERVICENAME --project=PROJECTNAME

Onboard a service

For describing the deployable Kubernetes resources of a service that gets onboarded, Helm charts are used with the following requirements:

1. The Helm chart has to contain exactly one deployment. In this deployment, the properties spec.selector.matchLabels.app and spec.template.metadata.labels.app have to be set.

2. The Helm chart has to contain exactly one service. In this service, the property spec.selector.app has to be set.

3. The Helm chart has to contain a values.yaml file with at least the image and replicaCount parameter for the deployment. These image and replicaCount parameters have to be used in the deployment. An example is shown below:

image: docker.io/keptnexamples/carts:0.11.1
replicaCount: 1

--- 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: your_deployment
spec:
  replicas: {{ .Values.replicaCount }}
  selector:
    matchLabels:
      app: your_service
  template:
    metadata: 
      labels:
        app: your_service
    spec:
      containers:
      - name: your_service
        image: "{{ .Values.image }}"

Note: If you are using custom configurations and you would like to have the environment variables KEPTN_PROJECT, KEPTN_STAGE, and KEPTN_SERVICE within your service, add the following environment variables to your deployment configuration.

env:
...
- name: KEPTN_PROJECT
  value: "{{ .Chart.Name }}"
- name: KEPTN_STAGE
  valueFrom:
    fieldRef:
      fieldPath: "metadata.namespace"
- name: KEPTN_SERVICE
  value: "{{ .Values.SERVICE_PLACEHOLDER_C.service.name }}"

Onboard a service:

• To onboard a service, use the keptn onboard service command and provide the service name, project name (--project flag), and the root directory of a Helm chart (--chart flag/):

keptn onboard service SERVICENAME --project=PROJECTNAME --chart=FILEPATH

• If you have already an archived Helm chart, the archive with ending .tgz can be referenced. In this case, the Helm chart will be stored unpacked.

keptn onboard service SERVICENAME --project=PROJECTNAME --chart=HELM_CHART.tgz

• Optional: In the case of using an archived Helm chart, you can achieve the same using keptn create service and keptn add-resource … –all-stages, e.g.:

keptn create service SERVICENAME --project=PROJECTNAME
keptn add-resource --project=PROJECTNAME --service=SERVICENAME --all-stages --resource=HELM_CHART.tgz --resourceUri=helm/SERVICENAME.tgz