I have always found Fiddler to be far more useful when debugging web traffic. It runs independently of applications, tabs, etc and allows for easy re-invocation of requests. And you can also test error handling easily by overriding the request and sending back an invalid response.
Fiddler is not the most user-friendly (it appears to be a bit thrown together), but it is very useful if you know how to find the right knobs.
I ran into this last night, Chrome shows me what requests are going over websockets, FF does not.