The order of the environment variables is atm crucial for the
run uc py.

Run uc py still used statefullSet instead of deployment

Fix #70

-Use aggregation from titan with dataflow
-Add an aggregation-feedback topic
-Enable logging in UC2 with new slf4j version
-Use newer Kafkastreams version
-Remove unused ConfigurationKeys class

Use the API instead of the kubectl program from the host machine.
This enables better portability of the script to other machines.

Apply the DRY principle and create method to load the methods with
a given path.
Further add comments to methods for better understanding and enable
the logging.

Use the kubernetes api in python to run the kubernetes creation
operations instead of the default platform commands.

Use the kubernetes api in python to run the kubernetes creation
operations instead of the default platform commands.

Add the hierachie structure to the kubernetes workload generator
files.
Use the new structure in the run uc scripts.

The pathes for the kustomize patch files were wrong in the shell
scripts and therefore running was not correct.
Further the execution time of the usecases should by calculated
because the m flag for sleep is not supported on macOS.

The environment variables in the kubernetes files need to be a
string in order to be computed by kubernetes.