This is a stub section, waiting for the actual documentation to be written
Syskit's underlying design is meant to allow for rich runtime monitoring and error handling.
- The goal of the main Syskit data structure is to track not only what is running, but also why, that is the relationships between the different pieces of software and hardware, which then allows to encode how some parts of the system influence other parts. When things "go wrong", it provides a rich context to understand what impact the error(s) have on the system.
- The ability of Syskit to "switch" between modes also provides a way to react to errors by switching operational modes, in a way other coordination systems do not have.
- The integration of "execution agents", that is the underlying OS processes within the execution model allows Syskit to monitor them and integrate them in the error detection and handling scheme
To be written:
- default error handling behavior
- exceptions and exception handling
- fault detection tables
- fault handling tables