Since I started logging exceptions on a production site I’m noticing a lot of them, especially 404s, more than I would expect for a site with barely any traffic, and I’d like to get to the bottom of whether they’re genuine users or just bots. To help with this, I want to capture the URL that the user was trying to visit before being redirected to the 404 route, so I can keep track of which non-existent routes are being mistakenly hit. I think I’m correct in assuming this URL should be available in the request, and that I just need to store the request and pass it through to the exception.
What’s the best way to do this in Laravel 8 onwards?
2
Answers
EDIT: just a few hours of using this solution in production, with the benefit of the newly-added request URLs, has confirmed for me just how many of these HTTP 4xx errors are junk - mostly automated bots and maybe a few script kiddies trying common routes. For this reason I've added some logic to ignore 404 and 405 errors, and may still add others that mostly contribute noise to the logfile.
This was harder than it should have been, but this is the solution I'm currently using to log the request with all exceptions. It's probably not the cleanest way to do it, but it works perfectly for my needs. Thanks to John Lobo's answer for pointing me in the right direction.
It works by inspecting each instance of the Exception class and using PHP's
instanceof
to check whether it's a HTTP exception or not. If it is, it gets logged with the request URL and returns a view with a status code. If it's a generic non-HTTP exception, it gets logged with the request URL and returns another view with no status code (or you can keep the default exception behaviour by removing the return block, which renders a blank screen in production).Catch
404
exceptions inHandler(AppExceptionsHandler)
.If you see Rendering Exceptions
so in the register method,call
renderable
Also, don’t forget to import and use