I’m struggling to set up a batch data processing architecture with Spark, Docker, and Kubernetes, as shown in the attached diagram. The examples online are unclear and often don’t work.