Programs crash. So far, so well-known. But how can we learn from our mistakes and use the errors to learn and improve? In this talk, a short overview of what is still possible when an application crashes is given. The talk is custom-tailored for embedded devices. Traditional signal handlers, dumps and debugging is explained, and it reveals how to use a combination of Yocto, Google Breakpad, and Sentry to produce, enhance, and collect crash information to our needs.