Page Load Timing — Full Resource Waterfall From a Real Browser
This loads a URL in a real browser rather than fetching it, and records every subsequent request the page makes: scripts, stylesheets, images, fonts, third-party calls, each with its own timing. The result is a waterfall showing what a page actually costs, which is invariably a great deal more than the document itself.
Why run it
From an infrastructure perspective the interesting content is the dependencies. A page that appears to be served entirely from your CDN usually turns out to depend on a dozen third-party hosts, each of which is an availability and latency risk you did not choose and cannot control. The waterfall shows them explicitly, along with which of your own assets are being fetched from origin rather than cache — a direct measure of whether your caching configuration is doing what you believe.
Why automate it
Page composition changes with every deploy, and it changes without anyone assessing the infrastructure consequences. A tag manager gains a new script, a marketing tool is added, an image is replaced with one an order of magnitude larger. Capturing the waterfall on a schedule and diffing it turns "the site got slower" into a specific list of resources that appeared, grew, or stopped being cached — which is the difference between a productive conversation and a week of speculation.
What it gives you
Visibility into what a page pulls in beyond your own infrastructure, and an attributable record of when that changed. It is the tool that turns a vague performance complaint into a named resource.