1. Install VVP (e.g. on minikube): https://docs.ververica.com/getting_started/installation.html
2. mvn clean package
3. (optional) Publish docker image e.g.

• docker build docker <YOUR_DOCKER_REGISTRY>/ververica/pyflink:1.11.2-scala_2.12
• docker push <YOUR_DOCKER_REGISTRY>/ververica/pyflink:1.11.2-scala_2.12

4. Add secrets for pulling docker images

• kubectl create secret docker-registry regcred --docker-server=<YOUR_DOCKER_REGISTRY> --docker-username=[gitlab-username] --docker-password=[token] --docker-email=[email] -n vvp-jobs

4. Create a deployment

• upload shaded jar in the UI (note: you might have to edit YAML manually for the first time to make it work)
• go to the YAML tab and add:

spec:
  template:
    spec:
      artifact:
        flinkImageRegistry: <YOUR_DOCKER_REGISTRY>
        flinkImageRepository: ververica/pyflink
        flinkImageTag: 1.11.2-scala_2.12
      kubernetes:
        pods:
          imagePullSecrets:
            - name: regcred

5. Job should finish successfully

In general: we package program with the jar and execute it on VVP. Alternative approach: use persistent volume.