The CPU performance view is one of the most important views in the system. It allows the breakdown of CPU-related events, multiple instances of the same machine events, processes affecting the CPU and more.
Use this view to monitor and identify performance issues that are directly related to CPU-intensive systems.
CPU – Part 1
| Legend | |
| 1 | CPU Usage – for the selected instance. This is the last sqlservr.exe process CPU for the given instance. |
| 2 | CPU Usage – for Machine. The total last CPU usage for the given machine. The default sample interval is every 60 seconds and can be changed in Settings. |
| 3 | CPU Usage Enhanced Chart. Shows all MSSQL instances for the same machine. Alert lines and alert indicators are associated with the “Machine SQL Process CPU Over Threshold” alert configuration type. |
| 4 | Server total CPU Usage Enhanced Chart. Follows the server’s total CPU over the selected timeframe. |
| 5 | Top processes. The AVG CPU time for the selected timeframe of the top OS processes. |
| 6 | Top cores. The AVG CPU usage for the selected timeframe for the machine’s cores. Core names as given by the OS. |
CPU – Part 2
| Legend | |
| 1 | CPU Waits – Chart of MSSQL wait events that are associated with CPU stress:
CXPACKET – Caused when parallelized query threads have an unequal workload, causing a block on the entire query until the longest-running thread is complete. See help for further discussion. SOS_SCHEDULER_YIELD – Caused when a task voluntarily yields the scheduler for other tasks to execute. During this wait, the task is waiting for its quantum to be renewed. CXCONSUMER – CXCONSUMER itself does not directly consume CPU resources. Instead, it represents a wait state where the Consumer thread is idle, waiting for data. However, high occurrences of CXCONSUMER can indicate that the system is heavily utilizing parallelism, which can lead to high CPU usage overall due to the large number of parallel threads being executed |
| 2 | Database CPU usage. This list is created directly from the statistics of the queries that are correlated with default databases. The CPU usage data is calculated as a ratio of the total. The sum of all values will always be 100%. This indicator will help recognize the most active databases. |
| 3 | Queries Grid – sorted by default by CPU |