ensure non-blocking execution

The current scripts and architecture is coded around certain assumptions of the system. It is possible, that some scripts could unexpectedly return a error code that would exit the entire script, even cleanup afterwards resulting in a messy filesystem. We should consider ways to ensure important steps like cleanup are still possible, even on a failing command.

For that we need to consider
- what are the most important general steps to always perform no matter the execution state (like cleanup step)
- how to implement handlers, catching and logging failures, without terminating
- which commands could lead to those errors (see curl fetching a 404 error page or grep not finding the directory and so on)