Application panics are unfortunate, but how you recover from an application panic is very important. Echo comes with a Recover middleware that is actually able to catch application panics using the Go recover built-n function to catch an application panic. This is possible because, if you recall, middleware functions effectively just wrap our handler functions. Since that is the case, the Recover middleware is able to do the following:

defer func() {
        if r := recover(); r != nil {
            //...
        }
    }()

With the preceding code in place, any application panic that befalls a running request handling goroutine will be caught. I strongly recommend reading the Go blog Defer, Panic and Recover for more information on how this built-in function works. The following diagram explains what Echo is able to do to catch panics and recover gracefully. If a panic is encountered within the handler function, or any nested middleware, the Recover middleware will catch the panic and return a generic error to the Echo framework, which will cause Echo to respond to the caller with the default error handler, which we discussed earlier in this chapter:

The consequence of not recovering from an application panic could be very embarrassing, as the consumers of your service will get a full stack trace of your application as the response. I would strongly recommend always wrapping your application in the Recover middleware just in case you have any kind of application panic within your service.