Complete Web Monitoring

Deciding How to Collect RUM Data

Having information on every user’s visit is tremendously useful, both for troubleshooting individual incidents and for determining whether your website is living up to its promises.

When deploying a RUM tool, your first decision is how to collect all this data. Your approach to collection significantly affects what you can do with your RUM solution, and there are advantages and disadvantages to each approach.

We’ve seen that there are two major ways of collecting data: on the server side and on the client side. Server-side collection approaches include server logging, reverse proxies, and inline sniffers or passive analyzers. Client-side collection approaches include desktop agents and JavaScript instrumentation in the browser.

Much of this collection technology resembles the monitoring approaches we looked at for a web analytics implementation, but it’s focused more on performance measurement. Consequently, passive analysis approaches are more common in RUM than they are in web analytics because they can collect network timing information and detect failed requests.

Server Logging

Web server logs give you only basic information about visitor performance. You’ll have the timestamps at which each object was requested, and basic data like what was requested.

You may have more advanced logging, either through specialized software or within the application server. This can tell you about key milestones of an object request, such as when the request was received, ...