Unlike many other monitoring services which just do ping or HEAD requests, we measure the full response time from the DNS lookup and opening the connection all the way through to loading all content and assets (JS, CSS, images, etc) from the full page. This gives you a more accurate response time as if a human visitor was loading the page in their browser.
The timer starts when we start the request and stops when the last element has fully downloaded. The maximum time the request will wait to complete is defined by the timeout value.
The dashboard shows the latest, real time response time. For round robin services this is the average across all monitoring node locations. For specific location services this is the lastest value.
The graphs show the actual response time value, not averaged.