The server is started standalone on a machine and listens on a dedicated port:
In the JSON file where the jobs are defined, you see the attributes of a job and you get an idea of what things I currently process. You can set a dependent job, the log level, the parameters for a job and interval how often a check should be done to run a job, etc. The process also creates log files for the job runs.
Hope this helpds to understand the way I am going. I will do some futher development and then run some more realistic examples, that have many ETL processes to see if it works out/behaves well. Scaling and performance will also be checked and running multiple servers.
Carpe Diem.