An often overlooked component of successfully implemented Microsoft Office SharePoint Server 2007 servers is performance monitoring. Performance monitoring should be an integral part of routine operations on every system-not just for Office SharePoint Server 2007 servers. However, this chapter will focus on performance monitoring for SharePoint Server 2007.
Ever-growing workload demands, shrinking schedules, and change of scope can cause server performance monitoring to be overlooked or completely ignored. Eventually, all complex systems will exhibit slowdowns of one severity or another for myriad reasons. This chapter will begin by identifying these potential bottlenecks and then give focused examples of how to resolve existing performance issues.
Of course, a contented SharePoint administrator is one whose server farm is operating well, with services that are meeting or exceeding the requirements of the business-and doing so with the least effort possible. The second part of this chapter will help SharePoint administrators get a little closer to that desired state by introducing Microsoft Operations Manager 2005-a product that can automatically monitor and report on the performance counters, event logs, and the health and availability of enterprise products and services from across the network, including SharePoint Server 2007. Monitoring system health closely is essential for spotting and addressing possible issues before they arise, quickly remedying problems that do occur, and understanding usage demands on the system to help predict future growth requirements.
The tool that most systems administrators find the easiest to use is the Microsoft Windows Performance tool (%systemroot%\system32\perfmon.exe). In previous versions of Windows Server, this tool was called Performance Monitor, or PerfMon. The default view of the Performance tool is called System Monitor. The Performance tool is always available, no matter which Windows Server system you work on. The Performance tool provides the required tools to diagnose virtually all performance issues on a Microsoft Windows Server system, regardless of the applications or products in use, with surprisingly little overhead. The Performance tool has the capability to monitor thousands of different counters, but it is best to identify possible bottlenecks and monitor only those. Trying to monitor too much will yield massive logs and essentially render the tool useless. For a potentially complex application, it is essential that the scope of monitoring be narrowed to effectively use the Performance tool.