SQL Engine Test
The SQL Engine test reports statistics related to the Microsoft SQL server engine.
Target of the test : A Microsoft SQL server
Agent deploying the test : An internal agent
Outputs of the test : One set of results for every Microsoft SQL server monitored
|
Measurement | Description | Measurement Unit | Interpretation |
---|---|---|---|
Cpu usage: |
The percentage of time for which the server’s CPU was engaged in processing requests to the server |
Percent |
A high value of this measure indicates a heavy load on the server. If this value comes close to 100%, it could indicate a probable delay in the processing of subsequent requests to the server. The detailed diagnosis measures associated with the Background processes measure of the MsSqlSysProcesses test will help you identify the processes that are consuming excessive CPU resources. |
I/O usage: |
The percentage of time for which the server was engaged in performing input/output operations |
Percent |
|
CPU idle time: |
The percentage of time for which the server was idle |
Percent |
A low value of this measure is indicative of high CPU utilization. |
Packets received: |
The rate at which input packets were read by the SQL server |
Pkts/Sec |
This measure is an indicator of the traffic to the server. |
Packets sent: |
The rate at which output packets were read by the SQL server |
Pkts/Sec |
This measure is an indicator of traffic from the server. |
Packet errors: |
The rate at which packet errors occurred |
Errors/Sec |
Ideally, this value should be 0. |
Disk reads: |
The rate of disk reads performed by the Microsoft SQL server |
Reads/Sec |
The value of this measure should be kept at a minimum, as disk reads are expensive operations. Ideally, data reads should be performed from the server cache and not directly from the disk. To ensure effective cache usage, allocate adequate memory to the Microsoft SQL server. |
Disk writes: |
The rate of disk writes performed by the Microsoft SQL server |
Writes/Sec |
The value of this measure should be kept at a minimum, as disk writes are expensive operations. Ideally, data should be written to the data cache and not directly to the disk. To ensure effective cache usage, allocate adequate memory to the Microsoft SQL server. |
Disk I/O errors: |
The rate of errors encountered by the Microsoft SQL server while reading and writing |
Errors/Sec |
Disk read/write errors are normally caused by the following reasons:
|
Note:
Sometimes, this test may suddenly stop reporting values for the CPU usage, I/O usage, and CPU idle time measures. In such a scenario, search the agent error log for the "msg 8115 Arithmetic overflow error". This is a Microsoft error, which occurs if the Microsoft SQL server being monitored is not restarted for 49 consecutive days or more. For more details about this error, refer to the following links:
https://docs.microsoft.com/en-us/sql/t-sql/functions/cpu-busy-transact-sql?redirectedfrom=MSDN&view=sql-server-ver15
https://docs.microsoft.com/en-us/sql/t-sql/functions/io-busy-transact-sql?redirectedfrom=MSDN&view=sql-server-ver15
If you find the "msg 8115 Arithmetic overflow error"in the agent error log, then you may want to consider restarting the target server during maintenance.