Hello beam community, I have a question arround beam python, and kafka and flink runner. As the kafka source need java, right now, the flink runner will try to start a docker container. I have trouble finding documentation arround this behavior. As we run flink in kubernetes or in docker environment, it will be pretty difficult to do docker in docker.
1_ Is it possible to run this part in a process if the dependency are installed on the flink runner? 2_ Is it possible to configure to run a worker pool like for python? (when using help there is no worker pool like in python) 3_ Is it possible to configure the host of the provision endpoints (as localhost is hard to configure in docker inside docker, or any options to configure the docker commands)? Thank you a lot for your help. Best regards
