Handle failures during progress parsing more gracefully

I propose an option that allows progress parsing failures to be ignored. This way progress parsing can continue in the event of a one-time error that would have otherwise interrupted the parsing thread.

In #302 the progress parsing fails because of an unexpected input ("N/A" instead of a formatted timestamp) in the first progress event. Although #315 fixes this, there can arise other circumstances where parsing can fail. I propose a general solution, to allow these failures to be ignored altogether. In most scenarios the consumer doesn't care if every once in a while an event gets dropped as long as they get notified of the overall progress.