Whether you're a network architect, a web developer, or a web administrator there's one tool that's a must have in your troubleshooting toolbox: a protocol analyzer.
Like many network focused folks, I traditionally rely upon ethereal (now Wireshark) for protocol analysis. It decodes just about every protocol up and down the stack, and it can import/export to a variety of formats. But being connected to the corporate LAN via an SSL VPN, wireshark is often constrained by it's own architecture. Because it inserts itself into the network stack to gather data, it can't decrypt the SSL encrypted packets, which makes it very difficult to follow an HTTP stream.
Additionally, unless you've got ninja-like skills and can write an appropriate filter to see just the data you're interested in, the signal-to-noise ratio in Wireshark can be overwhelming.
Many of our field engineers use browser plug-ins instead of Wireshark (that's going to take some time to get used to) for troubleshooting and verifying configuration of our products. Because BIG-IP is a full-proxy, and therefore doing all the communication with the client, it's important at times to peek under the covers of an HTTP request and check out things like cookies, headers, and the payload of request-response pairs.
After becoming curious about the traffic generated by Flowgram (how much, what type) yesterday I happened to stumble upon a new Firefox add-on called HttpFox. Most folks would suggest using Httpwatch and I have in the past, but after installing and using HttpFox, I'm not going back.
HttpFox can easily be launched via its icon in the status bar of Firefox with a single click. It docks itself on the bottom half of the browser, though you can easily manipulate the panels and change its size to suit your needs.
Once opened you can start analyzing a page by hitting the start button. URL requests start flying by (especially on Web 2.0 sites) automatically, though you can control the scrolling using the "autoscroll" option.
You can view associated headers, cookies, query strings, and post data for any individual request simply by clicking on the URL you want to examine. Filtering is available so you can restrict the list of URLs by type, file extension, host name, or whatever you'd like.
When examining post data you can choose to view it "raw" or "pretty", and content can also be viewed as the raw HTML received. This latter option is a nice alternative to the "view source" option in the browser.
Regardless of why you want to dig under the hood, HttpFox is a great way to do it. It won't let you change or manipulate the requests and data like some add-ons, but it will let you see all the nitty-gritty details about them in a well-designed interface that's easy to open, close, and manipulate.