Design – Build – Run: Applied Practices and Principles for Production-Ready Software Development

9.4. Analytics Overview

Various products (for example, WebTrends) analyze web server logs and provide reports and analysis. The packages essentially trawl the logs and capture a wide variety of information, including but not limited to:

• Pages visited

• Path taken to particular pages

• Time spent on each page

• Page performance

• Pages not visited

• Internal/external visits

• The pages visitors arrived from

Because these tools are analyzing the web logs, they may require specific logging configuration on the web server. For instance, the analytics software may require the logs to be in a specific format and/or contain specific information not included by default. It is important to understand the configuration requirements of the tools. In Internet Information Services, many different properties can be recorded in the logs (if logging is switched on). The following lists just a handful of the properties that can be output:

• Date and time

• Client IP address

• Website name

• Page name

• Query string

• Bytes sent

• Bytes received

Turning on additional logging could affect the performance characteristics of the online application and should therefore be understood prior to technical testing.

Some web analytics solutions are implemented using the master/agent paradigm. The agents gather data from the web server logs, analyze them, and then send data to the master for storage and display. It is important to understand the implications on performance and network traffic.

In addition, some web analytics software packages ...