In the preceding section, we started to think about fault tolerance in our frontend code. That is, our application needs to survive the loss of a failed component—at least in the short term. But what if there are certain kinds of errors that we can recover from? So instead of shutting down the component after detecting the error, we would take some alternative course of action; one that would still satisfy the user.

In this section, we'll look at the various ways our components can recover from failed operations. For example, we can retry an operation, or we could flush out the bad state of a component by restarting it. Sometimes, it makes sense to get input from the user on how they wish to proceed during a recovery effort.