Create resources for codefresh-guestbook

Before you can create an application in Codefresh, you need to create the resources used by the application:

  1. Install Argo Rollouts on the target cluster to which you will deploy the application
  2. Rollout resource defining the deployment strategy
  3. Service resource to expose the application to external traffic
  4. Analysis Template resource defining the validation requirements before deployment

Install Argo Rollouts on the target cluster

To apply the Rollout resource for the application, you must have Argo Rollouts installed on the target cluster. If not installed, follow the steps to install Argo Rollouts.

  1. In the Codefresh UI, go to Runtimes.
  2. Select Topology View.
  3. Select the target cluster, and then select + Install Argo Rollouts.

Install Argo Rollouts on target cluster

Install Argo Rollouts on target cluster

Create folder in Git for application resources

Create a folder in the Git repo in which to save all the resources.

  • In your Git repo, create a folder to store the resources needed to deploy the application.

    For example, /quick-start/

Create rollout.yaml

Create a rollout resource for the application you want to deploy.

To leverage Argo Rollout’s deployment capabilities, Codefresh are using the Argo’s Rollout resource instead of the native Kubernetes Deployment object. For detailed information on the fields you can define, see Argo Rollout specification.

  • In the Git repository create the rollout.yaml file, as in the example below.
kind: Rollout
  name: codefresh-guestbook-rollout
  replicas: 4
  revisionHistoryLimit: 2
      app: codefresh-guestbook
        app: codefresh-guestbook
        - image:
          name: codefresh-guestbook
            - name: http
              containerPort: 80
              protocol: TCP
  minReadySeconds: 30
          - templateName: background-analysis
        startingStep: 1
        - setWeight: 25
        - pause: {duration: 20s}
        - setWeight: 75
        - pause: {duration: 15s}

Fields in rollout.yaml

Rollout Field Notes
replicas When deployed, the rollout creates four replicas of the codefresh-guestbook application.
revisionHistoryLimit The number of replica sets to retain.
matchLabels The pods to select for this rollout. In Codefresh’s example, all pods with the label codefresh-guestbook are selected.
image The container image for the application with the version tag, in Codefresh’s example.
name The name of the application, codefresh-guestbook in Codefresh’s example.
canary The deployment strategy, canary meaning that the traffic is gradually routed to the new application. Starting with setWeight of 25% followed by a pause of 20 seconds, and the remaining 75% after verification.
templateName The analysis template used to validate the application metrics. Codefresh’s example has the background-analysis template, and interfaces with Prometheus to monitor and validate metric thresholds.

Create a service resource

Create a service resource to expose your application to external traffic.

  • Create a service.yaml resource for the application you want to deploy, as in the example below.

    Create it in the same folder in which you saved rollout.yaml.

apiVersion: v1
kind: Service
  name: codefresh-guestbook-svc
    - port: 8080
      targetPort: 80
    app: codefresh-guestbook # must be the same as the selector defined in rollouts.yaml
  type: LoadBalancer

Fields in service.yaml

Service field Notes
spec.ports The internal port, 8080 in Codefresh’s example, and external targetPort, 80 in Codefresh’s example. The pods to select, and MUST be identical to that defined in rollouts.yaml, codefresh-guestbook in Codefresh’s example.

Create an AnalysisTemplate for rollout validation

Create an AnalysisTemplate resource to validate that your changes conform to the requirements before deployment. This is the final resource you need before you can create the application.

The name of the AnalysisTemplate in the quick start example is background-analysis. The template interfaces with Prometheus as the third-party metric provider to validate metrics.

You can use any third-party metric provider supported by Argo Rollouts, such as Prometheus, Datadog, Wavefront, and more. Read the official documentation on Analysis section in Argo Rollouts.

  • In the Git repository create the analysisTemplate.yaml file, as in the example below.
kind: AnalysisTemplate
  name: background-analysis
    - name: prometheus-metric
      count: 4
      interval: 5s
      successCondition: result[0] >= 100
      failureLimit: 1
          query: |

Fields in analysisTemplate.yaml

Analysis Template field Notes
count The total number of measurements taken, 4 in Codefresh’s example.
interval The interval between measurement samplings, 5s in Codefresh’s example.
successCondition The requirement for the rollout to be considered a success. In Codefresh’s example, the resulting metric value must be equal to or greater than 100.
failureLimit The maximum number of failures permitted, 1 in Codefresh’s example. If the metric value is below 100 more than once, the rollout is aborted.
query The query submitted to the Prometheus server.

You are now ready to create the application in Codefresh.

What to do next

Create the codefresh-guestbook application