Traditionally, user-focused Web-application monitoring has been restricted to static log analysis or other forms of Web analytics. TeaLeaf Technology, a pioneer in user-based transaction monitoring since 1999, challenges this notion with its RealiTea platform. While not covered in a hands-on test, our interactions with the RealiTea platform through extensive demonstrations from TeaLeaf suggest it is a mature product that provides a superset of the user activity-monitoring and debugging features found in the xFire platform we tested.

The RealiTea platform takes the stance that what the user sees is the most important consideration. User session data is collected in a passive manner via a capture appliance attached to the spanning port of a switch or load balancer. Not having to install collection agents on origin Web servers reduces the chance for any performance and stability impact, and actually makes the system easier to deploy in large-scale environments.

The packet capture and reassembly of user sessions that RealiTea completes is more complex than simply monitoring HTTP sessions directly on a server. But that complexity provides the ability to get timing information regarding end-user performance by looking at TCP acknowledgments. This capability is quite valuable when trying to determine exactly what the end user experienced.

To properly debug any software application - Web-based or not - the problem must be reproduced. Yet when it comes to Web applications, simple log files often don't provide enough information, as you need to know what the user did and the context in which they did it. You need to know not only the series of requests that might have led up to a problem, but also any data a user might have entered, the type of browser they used, and the states of the server and the network at the time of the request.

After initial data capture, information is fed to a Windows 2000/2003 box running the RealiTea server software. During this stage, TeaLeaf allows filters to run against collected data because not all data can be useful for customer support, monitoring or debugging purposes. RealiTea can provide near real-time observation of site health and activity because captured data is initially aggregated and analyzed in a short-term memory database before it is moved to more long-term storage.

As data is collected, a Web browser can access a portal on a locally deployed RealiTea server to perform queries and monitor activity. The maturity of the platform is evident in the ease with which data can be searched and navigated. Drilling down into graphs, search results and alerts to a particular session, and then starting a session replay in the RealiTea Viewer is extremely natural.

The RealiTea Viewer session playback engine is very refined and clearly highlights user data entry and activity. Unfortunately because this viewer is also Internet Explorer-based it doesn't address the fact that browser-level issues can be important, and provides no way to playback sessions in a non-Internet Explorer browsers to address any problems related to client-side technology such as JavaScript.

The RealiTea platform also can track system data, for example, by collecting log data from servers. It also can capture application-level events through API calls that are inserted directly into Web application code. However, correlating this data with user activity for post-mortem application debugging purposes isn't necessarily the only use of the product, which also can be used for support and analytics.

While TeaLeaf's RealiTea is clearly a powerful platform, it comes with a price tag averaging about $150,000. For mission-critical sites with significant liability concerns or revenue at stake, this could be well worth it. However, for those with shrinking Web site budgets or less critical needs, the harsh reality is that the TeaLeaf platform, for all its power and polish, might not be in their future yet.

Originally published on Network World, Published: August 2, 2004.