Voxxed Days Vienna 2018

“Let’s throw an error” - developers are often quick with this decision as it makes life easier in that moment. But dumping failures to clients noticeable increases the overall system complexity. “Please try again in a few minutes”. Introducing long running behavior can drastically improve this situation. We will look at multiple interesting patterns of long running behavior in this talk. Services could for example react immediately but switch to asynchronous responses with persistent state when leaving the happy case. This allows to take seconds, minutes or even weeks to sort out failures which can involve waiting for services to become available or humans to intervene. Doing so, even if only for the failure case, makes everybody’s life easier. In this talk we want to iterate over various patterns by showing concrete examples. We will also give hints on proper implementations using hands-on source code examples.