What is a GDN Request?
Your GDN package includes a number of requests that the GDN can make to your website’s server. A ‘request’ is any HTTP request processed by the GDN that results in a response with HTML or XML resources (specifically text/html, application/xhtml+xml, text/xml or application/xml).
The following is not counted as a request:
- Requests that result in a 301, 302 and 304 response
- PDF, document downloads etc.
- JavaScript, JSON, or CSS
- Media files (images or video)
- or any other requests that do not include HTML or XML content
Therefore, the count of requests from a single page view can vary based on the number of resources the webpage needs to request from the server.
Analyzing Traffic to Estimate Requests
Web analytics tools, like Google Analytics, monitor user interactions with your website. These tools center around 'page views,' which usually track when a user opens a webpage. Yet, a single page often requires several HTTP requests to load all its content. Consequently, a single 'page view' in web analytics might correspond to multiple requests in the GDN.
Furthermore, web analytics tools gather data directly from the website as the page loads in the browser. In contrast, the GDN request count stems from network traffic analysis. Any human or bot can generate network traffic. It’s possible for end-user ad-blocking browser plugins to block the collection of web analytics data, leading to under-counting in the web analytics tool. On the other hand, it could also happen that a CDN in front of the GDN served cached pages to the end users, resulting in a lower count of GDN requests.
What is Bot Traffic?
Bot traffic refers to non-human visits to websites. It's not inherently good or bad; the nature depends on the bots' intent.
Good bots, like search engine crawlers or digital assistants (e.g., Siri, Alexa), serve useful purposes. Bad bots engage in malicious activities like credential stuffing, data scraping, or DDoS attacks. Even seemingly 'harmless' bad bots disrupt analytics and lead to cyberattacks.
Regardless of their intent, over 40% of internet traffic comprises bots, and a significant portion of your network traffic could be attributed to bots.
Identifying Bot Traffic
In Google Analytics, traffic from known bots and spiders is automatically excluded, which means that the number of registered page views may be lower than the actual number of GDN requests.
To identify bot traffic, your IT ops team could analyze network requests, or you can use integrated tools like Google Analytics for the following indications:
- Unusually high pageviews: a sudden, abnormal surge in pageviews might indicate bot activity.
- Unusually high bounce rates: unexpected spikes suggest bots directed to specific pages.
- Odd session duration: bots may browse slowly (high duration) or quickly (low duration).
- Fake conversions: spam form submissions from bots.
- Unanticipated traffic origin: surges from unlikely regions may hint at bots.
Managing Bot Traffic
- Robots.txt file: instructs bots to stay away from certain webpages. However, only good bots comply; malicious bots disregard.
- Bot management solution: your website’s CDN may support user agent filtering for bot management, such as Cloudflare Bot Management.
Estimating GDN Requests
For most websites, the minimum package of requests is enough. However, for very high traffic sites, it can be very challenging to forecast. Elements to consider when estimating requests include but are limited to:
- Your website's total page views (from your web analytics tool)
- Each localized site’s expected traffic (from your estimated user base per locale)
- A buffer for growth as you expand globally
- Other network traffic
GDN Request Analytics
To prevent overages, you can monitor your GDN Request usage in the Site Traffic Report. This report can also be emailed to you or any stakeholder.