fix pipeline crash because of API outage

After retry-timeout, an exception is raised and that is not caught.

So it seems that the API is expected to be returning 2xx responses before retry-timeout.

Increasing the retry timeout would help, but not if the outage is longer. In that case, the guests would still be orphaned, RepeatTimer catching the exception and terminating the pipeline.

I’ve made handle_pipeline_cancellation catch the exception and continue, but maybe RepeatTimer should also not just break and die - at least not without outputting something to the log.

fix: catch exception in handle_pipeline_cancellation also: do not break in the RepeatTimer - may spam the log with messages. What are any other side-affects? Please consider. Thanks!

Closes TFT-4698