workflow example

workflow example: new feature in kubewarden-controller + helm charts

This is a walkthrough on how I completed the changes in PR https://github.com/kubewarden/kubewarden-controller/pull/304, which included:

• kubewarden-controller: bump go 1.19, refactor of opentelemetry export code.
• helm-charts: fixes, and add tls.insecure
• docs
• tests: deploy Kubewarden stack + opentelemetry stack, test by hand

Needed tasks

• 1. Changes to controller repo

  $ cd /home/vic/suse/kw/kubewarden-controller
  $ git pull
  $ git checkout -b bump-controller
  
  # loop of code changes, commit them, etc
  $ make build
  $ make test

• 2a. Create container image

  # create container image
  $ IMG=controller:viccuad-opentelemetry make docker-build
  $ docker tag controller:viccuad-opentelemetry ghcr.io/viccuad/test/controller:viccuad-opentelemetry
  $ docker push ghcr.io/viccuad/test/controller:viccuad-opentelemetry

• 2b. Run controller binary outside of the cluster against a cluster Follow https://github.com/kubewarden/kubewarden-controller/blob/main/CONTRIBUTING.md#running. In this case, with opentelemetry, it's not worth it to expose the opentelemetry-collector outside of the cluster, to try and make the controller outside of the cluster connect to it.

3. Changes to helm chart kubewarden-controller This ones: https://github.com/kubewarden/helm-charts/pull/136

   $ cd /home/vic/suse/kw/helm-charts
   $ git pull
   $ git checkout -b opentemeletry

   # loop of code changes, commit them, etc
   
   make generate-values
   make check-generated-values
   

4. Deploy kubewarden stack + opentelemetry stack Following instructions from https://github.com/kubewarden/kubewarden-controller/pull/304#issue-1391142647, I deploy the helm charts from the helm-charts/charts/ folder to test my changes.

   Apart from the new values that are added , I need to configure the helm charts to test the relevant parts, with a values.yaml:

   ---
   # kubewarden-controller-opentelemtry.yaml
   telemetry:
     enabled: True
     metrics:
       port: 8080
     tracing:
       jaeger:
         endpoint: "my-open-telemetry-collector.jaeger.svc.cluster.local:14250"
         tls: # new 
           insecure: true # new

   ---
   # my-controller-image.yaml
   image:
     repository: ghcr.io/viccuad/test/controller
     tag: viccuad-opentelemetry

   ---
   # kubewarden-defaults-opentelemtry.yaml
   policyServer:
     telemetry:
       enabled: True

   I deploy cert-manager, jaeger, opentelemetry, configure their resources as listed in that comment.

   I deploy the kubewarden-controller chart by passing the 2 values files with --values kubewarden-controlleropentelemtry.yaml --values my-controller-image.yaml. Both files could clash on setting values, and would need to be merged, but not on this case.

   I deploy kubewarden-defaults with --values kubewarden-defaults-opentelemetry.yaml.

5. Test by hand. Or run specific e2e tests, that I would have just written for this feature.

6. Push branches to my remotes on GH, open PRs, iterate on feedback