eG Monitoring
 

Measures reported by MsSqlWaitTypeTest

In SQL Server, wait types represent the discrete steps in query processing, where a query waits for resources as the instance completes the request. By analysing wait types and their wait times, administrators can receive quick and objective evidence of performance bottlenecks and their probable causes.

This test reports the number, nature, and duration of waits, thereby leading you to the specific wait types that may have contributed to a general slowdown/deterioration in server performance.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Waiting_tasks_count Indicates the number of waits of this type during the last measurement period. Number This counter is incremented at the start of each wait.
Avg_wait_time_ms Indicates the total wait time for this wait type during the last measurement period. MilliSecs A low value is desired for this measure.

When a user complaints that query execution takes too long, you can compare the value of this measure across wait types to know which type of wait is the key contributor to delays in query processing.
Avg_signal_wait_time_ms Indicates the total signal wait time for this wait type during the last measurement period. MilliSecs The signal wait is the time between when a worker has been granted access to the resource and the time it gets scheduled on the CPU. A high value for this measure may imply a high CPU contention. To know which wait type registered the highest signal wait time and probably caused the CPU contention, compare the value of this measure across wait types.
Avg_res_wait_time_ms Indicates the total resource wait time for this wait type during the last measurement period. MilliSecs Resource wait time is the actual time a worker waited for the resource to be available. A high value for this measure indicates a delay in acquiring a resource. To know which wait type waited the longest for a resource and therefore contributed to a server slowdown, compare the value of this measure across wait types.
Per_wait_time Indicates the percentage of total wait time (across wait types) during which wait events of this type occurred. Percent When a user complaints that query execution takes too long, you can compare the value of this measure across wait types to know which type of wait is the key contributor to delays in query processing.
Per_signal_wait_time Indicates the percentage of total signal wait time (across wait types) during which wait events of this type waited for a signal. Percent The signal wait is the time between when a worker has been granted access to the resource and the time it gets scheduled on the CPU. A high value for this measure may imply a high CPU contention. To know which wait type registered the highest signal wait time and probably caused the CPU contention, compare the value of this measure across wait types.
Per_resource_wait_time Indicates the percentage of total resource wait time (across wait types) during which wait events of this type waited for a resource. Percent Resource wait time is the actual time a worker waited for the resource to be available. A high value for this measure indicates a delay in acquiring a resource. To know which wait type waited the longest for a resource and therefore contributed to a server slowdown, compare the value of this measure across wait types.