The CPU - Processor Queue Length alarm becomes active when the number of Windows threads waiting for CPU resources exceeds a threshold. This value is taken over a specific number of background collections. Sustained high processor queue length is a good indicator that you have a CPU bottleneck.
Note that since this is a Windows-based alarm, it can include CPU that was consumed by non-SQL Server processes.
TIP: When monitoring the Spotlight Diagnostic Server host, the Processor Queue Length shown here can be artificially high, especially on single processor machines. Since Spotlight on SQL Server is a multi-threaded application, the process of running its data collection queries causes extra Spotlight client threads to be scheduled for execution, resulting in a higher than normal Processor Queue Length figure and the raising of this alarm.
When the alarm is raised
If SQL Server is consuming most of the CPU and your system supports many concurrent users, you might benefit from SQL Server's lightweight pooling option, which causes SQL Server sessions to be scheduled as fibers rather than threads. On some heavily-loaded systems, this can save a small amount of CPU. Ensure you test this properly, because on some systems it can actually increase SQL Server's CPU demands.
This option can be changed by using the SQL Server | Configuration Drilldown. Changes to this option do not take effect until the SQL Server instance is stopped and restarted.