| TABLE 1: Counters Used in the Test |
| Object Name |
Counter |
Definition/Notes |
| Processor |
% Processor Time |
Indicates the utilization of each CPU. |
| Memory |
Pages/Sec |
Measures pages per second paged out of RAM to disk or into RAM from disk. Excessive paging causes more I/O overhead. |
| Physical Disk |
Average Disk Queue Length |
When the queue length exceeds 2 read or write requests per drive in the array, you have an I/O bottleneck. This counter measures I/O waits when the load on the server is heavy. |
| SQL Server Buffer Manager |
Buffer Cache Hit Ratio |
Indicates how often SQL Server uses the SQL Server Buffer Cache. This number should be as close to 100 percent as possible. |
| SQL Server Latches |
Average Latch Wait Time (milliseconds) |
Latches ensure that no other task can modify the contents of a row or page structure while that row is being transferred from the storage engine to the relational engine. High latch counts per second suggest I/O bottlenecks. |