You can set up your own environment, which will allow you to connect to Ververica/K8S operator. You need to fill properties within config/<env_name> directory. config/flink_deployment.yml is default deployment descriptor file for each environment (you can overwrite it using --file-descriptor-path flag). You can use jinja for templating (look at the tests for an example). base environment is the default environment. Others environment override parameters from base. You need to have in base or your own environment profile.yml file with given schema:

deployment_mode: <VVP|K8S_OPERATOR>
docker_registry_url: <docker_registry_url>

You need also extra file(s) with Ververica/Kubernetes configuration and optionally other configuration used for templating.

Example vvp.yml:

vvp:
  url: <ververica_url>
  namespace: <ververica_namespace>
  deployment_target: <some_deployment_target>

Example k8s.yml:

k8s:
  namespace: test_ns

For most of the command, you can pass the environment name with --env or export environmental variable as SCLI_ENV.

scli project deploy \
  --vvp-api-token "x§11d091jd1jd9jasd0j" \
  --docker-image-tag "${CI_COMMIT_TAG}" \
  --profile "dev"

scli project build \
  --docker-image-tag "latest"

scli docker login \
  --username "user" \
  --password "password" \
  --profile "dev"

Parameters--docker-image-tag is optional and has default value latest.

Scli uses requests library to deploy jobs to Ververica Platform via REST Api. Currently requests does not support automatic downloading of intermediate certificates so entire chain of certificates should be present before making a http call

example:

sh -c 'CA_CERT_PATH=`python3 -c "import requests; print(requests.certs.where())"`;for CERT in certs/*; do cat ${CERT}; done >> ${CA_CERT_PATH}';

• pipenv

• make setup - Install dependencies required to build a wheel package
• make cli/package - Create a wheel package

• make cli/install - Install a wheel package
• make cli/install/force - Reinstall a wheel package