Identify resources loaded from the network

Jeff Posnick
Jeff Posnick

The Network panel in your browser's DevTools helps identify what resources are loaded and when they are loaded. Each row in the Network panel corresponds to a specific URL that your web app has loaded.

Chrome DevTools' network panel.

This guide uses screenshots and descriptions based on Chrome's DevTools interface. Other browsers support similar functionality, but the overall interface will be different if you're not using Chrome.

Know what you load

Coming up with the right caching strategies for your web application requires getting a handle on what exactly you're loading. When building a reliable web application, the network can be subject to all kinds of dark forces. You need to understand the network's vulnerabilities if you hope to cope with them in your app.

You might think that you already have a pretty good idea as to what your web application loads. If you're just using a small scattering of static HTML, JavaScript, CSS, and image files, that might be true. But as soon as you start mixing in third-party resources hosted on content delivery networks, using dynamic API requests and server-generated responses, the picture quickly becomes murkier.

A caching strategy that makes sense for a few small CSS files probably won't make sense for hundreds of large images, for instance.

Know when you load

Another part of the overall loading picture is when everything gets loaded.

Some requests to the network, such as the navigation request for your initial HTML, are made unconditionally as soon as a user visits a given URL. That HTML might contain hardcoded references to critical CSS or JavaScript files that must also load in order to display your interactive page. These requests all sit in your critical loading path. You will need to aggressively cache these to be reliably fast.

Other resources, such as API requests or lazy-loaded assets, might not start to load until well after all of the initial loading is complete. If those requests only happen following a specific sequence of user interactions, then a completely different set of resources might end up being requested across multiple visits to the same page. A less aggressive caching strategy is often appropriate for content that you've identified as being outside the critical loading path.

Advanced techniques outside the scope of this guide, like <link rel="preload">, add a twist to this story by giving a head start to what would otherwise be a late-loaded request.

The Name and Type columns help with the what

The Name and Type columns help provide a meaningful picture of what's being loaded. The answer to "what's loading?" in the example above is a total of four URLs, each representing a unique type of content.

Chrome DevTools' network panel showing four files loading.

The Name represents the URL that your browser requested—though you'll only see the last portion of the URL's path listed. For example, if https://example.com/main.css is loaded, you'd only end up seeing main.css listed under Name.

The last few characters of the URL's path, following the period (e.g. "css"), are known as the URL's extension. The URL's extension generally tells you what type of resource is being requested, and maps directly to the information shown in the Type column. For example, v2.html is a document, and main.css is a stylesheet.

The Waterfall column helps with the when

Examine the Waterfall column, starting at the top and working your way down. The length of each bar represents the total amount of time that was spent loading each resource. How can you tell the difference between a request that's made as part of the critical loading path and a request that's fired off dynamically, long after the page's initial load is complete?

The first request in the waterfall is always going to be for the HTML document, for example, v2.html. All of the subsequent requests will flow (like a waterfall!) from this initial navigation request, based on what images, scripts, and styles the HTML document references.

Chrome DevTools' waterfall view.

The waterfall shows that as soon as v2.html has finished loading, the requests for the assets it references (also referred to as subresources) start. The browser can request multiple subresources at the same time, and that's represented by the overlapping bars in the Waterfall column for main.css and logo.svg. Finally, you can see from the screenshot that main.js starts loading last, and it finishes loading after the other three URLs have completed as well.

Last updated: Improve article

Codelabs

See it in action
Learn more and put this guide into action.