Implement proper exception handling throughout the project

Over the past few weeks, we rushed to get features working on KCIv2 and consciously ignored some engineering best practices while doing so (which is OK and is understandable given the context).

At this point, we have frequent crashes due to not handling exceptions properly, e.g. https://github.com/kernelci/kernelci-pipeline/issues/468 and https://github.com/kernelci/kernelci-pipeline/issues/475 being only recent examples. This affects development efforts as such crashes require immediate attention to get the system up and running again, while delaying important results.

Exceptions should be properly handled in every part of the code that *could* fail (including handling `KeyError` for every dictionary access that isn't properly guarded by a type/key existence check).

Start by reviewing the crashes described in kernelci-pipeline issues 468 and 475, then inspect the project areas where similar unhandled failures can occur. The issue names no files, tests, or entry points, so the scope must be established before work begins. Done means exceptions are properly handled throughout the project, including unguarded dictionary access.

Written by the indexing model from the issue text.