About the Solution

[Previous] [Next]

At Microsoft, our most important project metrics center around bugs: the number of bugs opened today; the number of bugs resolved today; the number of bugs assigned to different subgroups; and the number of bugs per area, per severity, and per priority. We use these metrics to measure the overall health of the project, to determine whether it is still on schedule, and to project when we think the product will be ready for a beta or final release.

These metrics come from our bug tracking system—an internal tool that stores information in a Microsoft SQL Server database. Every week, our test leads prepare detailed reports with charts showing the bug count activity during the previous week. These reports are extremely useful; however, when we approached critical points in the Microsoft Office 2000 cycle, the program managers wanted to see some of these metrics daily. We decided to put the Office Web Components to the test and develop a solution that produced charts of these important metrics on a daily basis. We offered several on-demand charts that displayed the metrics' values the moment a client requested them. This solution was so successful that we began getting requests from many other product groups within Microsoft for early builds of the components as well as for instructions on how we generated the charts.

In this chapter, I will describe exactly how we built this solution, and I will describe some more advanced techniques. However, I will present this solution in a slightly different context. Because bug tracking is relevant primarily to software companies, I decided to develop a solution around a similar process that most large companies have: tracking helpdesk logs. A helpdesk is a group of employees or contractors who help others in the company solve their technical problems, which are often related to software or hardware. Any time a new problem is reported, a new log is generated (some companies refer to this as a ticket) and the log progresses through a series of states until it is closed. The manager of a helpdesk needs to monitor a number of metrics on a daily basis to measure how successful the helpdesk is. For example, the manager often wants to see the number of new logs opened that day and the number of logs currently active, broken down by technician or type of technology. The manager might also want to watch the current call volume to determine whether there are enough technicians to handle the incoming calls. This chapter's solution, the Helpdesk Reporting solution, shows how to implement the presentation of a set of metrics in a variety of ways: as a server-generated chart, as a client-side interactive chart, and as a real-time updating chart.