2.1.3 Examples of performance monitoring
- Organization of this subsection
(1) SAP system's response times
These examples monitor the SAP system's response times in order to check trends in performance throughout the entire SAP system.
(a) Records and fields related to response times
The following table lists and describes the records and fields that are related to response times.
|
Record to be used |
Field to be used |
How to interpret the value (example) |
|---|---|---|
|
PI or PI_DIA |
ResponseTime |
Average dialog step processing time |
|
DBRequestTime |
Average time required to process a logical database request |
|
|
QueueTime |
Average wait time in the dispatcher queue |
(b) How to monitor
- Monitoring the dialog response time
-
To monitor the dialog response time in the SAP system, you can use the Dialog ResponseTime alarm (field name: ResponseTime) that is provided as a monitoring template.
If ResponseTime is equal to or greater than the threshold value, performance of the entire SAP system might have become degraded. Determine the bottleneck by monitoring the load status of the entire SAP system and the database request time.
- Monitoring the database request time
-
To monitor the database request time, you can use the Dialog ResponseTime report (field name: DBRequestTime) that is provided as a monitoring template.
If the value of DBRequestTime is high (exceeds 40% of the value of ResponseTime - QueueTime field), a problem might have occurred in buffering at the application server, in optimization of SQL statements (ABAP), or in the database server.
- Monitoring the load status in the entire SAP system
-
To monitor the load status in the entire SAP system, you can use the Dialog ResponseTime report (field name: QueueTime) that is provided as a monitoring template.
If the value of QueueTime is high (exceeds 10% of the value of ResponseTime field), the workload in the entire SAP system might have become elevated.
(2) Monitoring SAP buffers
These examples monitor the SAP buffers in order to ensure that the SAP system is operating efficiently.
By using the SAP buffers efficiently, you can reduce the response times for applications that are executed frequently, such as repetitive jobs.
(a) Records and fields related to the SAP buffers
The following table lists and describes the records and fields that are related to the SAP buffers.
|
Record to be used |
Field to be used |
How to interpret the value (example) |
|---|---|---|
|
PI or PI_BUFF |
Program HitRatio % |
Percentage of queries that did not need to access the database because the program was in the program buffer (buffer hit rate) |
|
CUA HitRatio % |
Percentage of queries that did not need to access the database because the menu information was in the CUA buffer (buffer hit rate) |
|
|
GenericKey HitRatio % |
Percentage of queries that did not need to access the database because the table data (multiple records) was in the generic key buffer (buffer hit rate) |
|
|
SingleRecord HitRatio % |
Percentage of queries that did not need to access the database because the table data (1 record) was in the single record buffer (buffer hit rate) |
|
|
PI_BUFF |
Program Swap |
Buffer-full swap count in the program buffer per minute. We recommend setting a value of 0 to this field. |
|
CUA Swap |
Buffer-full swap count in the CUA buffer per minute. We recommend setting a value of 0 to this field. |
|
|
GenericKey Swap |
Buffer-full swap count in the generic key buffer per minute. We recommend setting a value of 0 to this field. |
|
|
SingleRecord Swap |
Buffer-full swap count in the single record buffer per minute. We recommend setting a value of 0 to this field. |
(b) How to monitor
- Monitoring the program buffer
-
You can monitor the program buffer hit rate and swap count.
To monitor the program buffer hit rate, you can use the SAP Buffer Hitratio report (field name: program HitRatio %) that is provided as a monitoring template. If this value is low (less than 80%), the number of user requests (other than for repetitive jobs) might have increased.
To monitor the program buffer swap count, you can use the SAP Buffer Hitratio report (field name: Program Swap) that is provided as a monitoring template. If this value is greater than 0, the size of the program buffer might be too small.
- Monitoring the buffer for menu information
-
You can monitor the buffer hit rate and swap count for menu information.
To monitor the buffer hit rate for menu information, you can use the SAP Buffer Hitratio report (field name: CUA HitRatio %) that is provided as a monitoring template. If this value is low (less than 80%), the volume of menu operations (other than for repetitive jobs) might have increased.
To monitor the swap count for menu information, you can use the SAP Buffer Hitratio report (field name: CUA Swap) that is provided as a monitoring template. If this value is greater than 0, the CUA buffer might be too small.
- Monitoring the GenericKey buffer for table data
-
You can monitor the GenericKey buffer hit rate and swap count for table data.
To monitor the buffer hit rate for table data, use the SAP Buffer Hitratio report (field name: GenericKey HitRatio %) that is provided in the monitoring templates. If this value is low (less than 80%), the generic key buffer might be too small or there might be a problem in the table allocation method for the generic key buffer.
To monitor the buffer swap count for table data, use the SAP Buffer Hitratio report (field name: GenericKey Swap) that is provided in the monitoring templates. If this value is greater than 0, the generic key buffer might be too small.
- Monitoring the SingleRecord buffer for table data
-
You can monitor the SingleRecord buffer hit rate and swap count for table data.
To monitor the buffer hit rate for table data, use the SAP Buffer Hitratio report (field name: SingleRecord HitRatio %) that is provided in the monitoring templates. If this value is low (less than 80%), the single record buffer might be too small or there might be a problem in the table allocation method for the generic key buffer.
To monitor the buffer swap count for table data, use the SAP Buffer Hitratio report (field name: SingleRecord Swap) that is provided in the monitoring templates. If this value is greater than 0, the single record buffer might be too small.
(3) Monitoring the SAP memory
This example monitors the SAP memory specific to an SAP system in order to check trends in performance degradation in the entire SAP system due to insufficient SAP memory area.
(a) Records and fields related to the SAP memory
The following table lists and describes the records and fields that are related to the SAP memory.
|
Record to be used |
Field to be used |
How to interpret the value (example) |
|---|---|---|
|
PI or PI_MEM |
EsAct % |
Current expansion memory use rate |
|
HeapAct % |
Current heap area use rate |
|
|
PrivWpNo |
Number of work processes placed in the PRIV mode |
|
|
R3PagingUsed % |
Paging area use rate |
|
|
R3RollUsed % |
Roll area use rate |
(b) How to monitor
- Monitoring the expansion memory use rate in the SAP memory
-
To monitor the expansion memory use rate in the SAP memory, you can use the Extended Memory alarm that is provided as a monitoring template.
If the Extended Memory alarm status is abnormal or a warning, there might not be enough extended memory area.
- Monitoring the heap area use rate in the SAP memory
-
To monitor the heap area use rate in the SAP memory, you can use the Heap Memory alarm that is provided as a monitoring template.
If the Heap Memory alarm status is abnormal or a warning, there might not be enough heap area, or a dialog work process might have resulted in a short dump.
- Monitoring the number of work processes that are in PRIV mode
-
To check the number of work processes that are in the PRIV mode, you can use the SAP Memory Detail drilldown report (field name: PrivWpNo) that is provided as a monitoring template.
If this value is 1 or greater, a dispatcher wait time might have increased.
- Monitoring the paging area use rate in the SAP memory
-
To monitor the paging area use rate in the SAP memory, you can use the Paging Area alarm that is provided as a monitoring template.
If the Paging Area alarm status is abnormal or a warning, there might not be enough paging area.
- Monitoring the roll area use rate in the SAP memory
-
To monitor the roll area use rate in the SAP memory, you can use the Roll Area alarm that is provided as a monitoring template.
If the Roll Area alarm status is abnormal or a warning, there might not be enough roll area, or a dialog work process might have been placed in PRIV mode.
(4) Monitoring SAP system logs and CCMS alerts
The SAP system outputs to the system log the events that have occurred and error information.
The SAP system is equipped with a Computer Center Management System (CCMS) that analyzes system operations management and loads.
PFM - Agent for Enterprise Applications can periodically output to a text file the system logs and alerts (alert information) that have occurred in CCMS's warning monitor.
You can use information in this text file to monitor the status of the SAP system by linking it to another program that monitors logs.
(a) How to monitor
- Monitoring system logs
-
The system log information extraction function enables you to periodically output to a text file the system log information that is specific to the events and failures that have occurred in the SAP system.
This function outputs the following information:
-
Time the message was recorded
-
Server that recorded the message
-
User that recorded the message
-
Program that recorded the message
-
Message number
-
Message
For details, see Chapter 5. Extracting System Log Information.
-
- Monitoring CCMS alert information
-
The CCMS alert information extraction function enables you to periodically output to a text file the warning events (alert information) that occur in CCMS's Alert Monitor.
This function outputs the following information:
-
Alert ID
-
ID of the MTE associated with the alert
-
Severity of the alert
-
General property
-
Message
For details, see Chapter 6. Extracting CCMS Alert Information.
-