Resiliency Is a Good Thing, Right?

A lot of systems could use something like that because you can't depend on third party services. I mean, if you depend on them, you're going to get a brutal wakeup call as soon as they have an issue of their own. So at some point your whole application in a Twitter is resilient, but that doesn't mean your Phoenix or electric application is resilient, right? Yeah, exactly. The processes and supervision trees are not the source of failure within the process. You don't even need external API dependencies. Like lixir, what are the solution for the caching problem? Just spinning up a process or ETS store which stores results of a failure in case it happens again.