At every transaction step, ABAP work processes first write statistics and application statistics data to a buffer. This data is then saved in file system files on the host of the relevant instance. Separate files of this type are created for the statistics and the application statistics every hour. The total number of files is usually restricted to 48 (profile parameters stat/max_files or stat/as_max_files). This data is therefore stored for two days before it is overwritten.
At the next level, this statistics data is aggregated and saved as daily aggregates (that is, with the time resolution of days) in the workload collector database. A corresponding report runs once an hour on every instance for this. You can define yourself the profiles for which statistics data is to be saved in the database. These profiles each cover a separate topic area and represent the existing Analysis Views of the Workload Monitor. In detail:
|
Profile |
Analysis View |
AA |
Application profile |
|
AB |
Application hitlist: DB accesses |
Hitlists → Top DB Accesses functions |
AC |
Application hitlist: response times |
Hitlists → Top Response Time functions |
EA |
Component hierarchy |
This data does not create a separate analysis view, but rather provides additional information for other analysis views. |
EB |
Orgunit |
|
WA |
Task types profile |
|
WB |
Time-dependent task types profile |
|
WC |
Time, program, task types profile |
|
WD |
Database procedures profile |
|
WF |
External Systems |
|
WG |
Function code profile |
|
WH |
Frontend profile |
|
WJ |
Memory usage profile |
|
WK |
Spool activity profile |
|
WL |
Table profile |
|
WM |
User, transaction profile |
Transaction profiles and User profile and accounting statistics → user profile |
'WN' |
User profile |
User profile and accounting statistics → Accounting statistics |
WO |
RFC client profile |
RFC profiles → RFC client profile |
WP |
RFC client destination profile |
RFC profiles → RFC client destination profile |
WQ |
RFC server profile |
RFC profiles → RFC client profile |
WR |
RFC server destination profile |
RFC profiles → RFC server destination profile |
WT |
Hitlist: database accesses |
Hitlists → Top DB accesses dialog steps |
WU |
Hitlist: response times |
Hitlists → Top response time dialog steps |
WV |
Spool profile |
Spool statistics (Overview tab page) |
WW |
HTTP client profile |
Web statistics → WEB client statistics |
WX |
HTTP client destination profile |
Web statistics → WEB client destination statistics |
WY |
Database connection |
The statistics data at file system level contains the statistics records as they are written by the kernel after each transaction step, while the data in the workload collector database is already aggregated, meaning, for example, that it is no longer possible to analyze parts of a transaction. This means that after two days, you can no longer perform an analysis below transaction level.
The daily aggregates are also further aggregated, in two different ways:
· The statistics data of all instances in a system is combined to a TOTAL instance, which you can display as an additional instance of a system in the workload monitor.
· The daily aggregates are combined into weekly and monthly aggregates.
You can also define in this case the profiles for which the weekly and monthly aggregates and the aggregates for the TOTAL instance are to be created.
By default, daily, weekly, monthly, and TOTAL aggregates are generated for all profiles. You can define the retention periods for all of these aggregates, broken down by profiles.
To change the generation and retention period of the statistics data, or to monitor the correct functioning of the corresponding collectors, start the workload monitor (transaction ST03) and ensure that you are in the Expert user mode.
· To determine the profiles for which daily aggregates are to be created in the workload collector database, choose the following path in the navigation bar: Collector and Performance DB → Workload Collector → Instance Collector → Control. You can only display data in the workload monitor for profiles that are activated in this table.
· To check that the corresponding collector is functioning correctly, choose the following path in the navigation bar: Collector and Performance DB → Workload Collector → Instance Collector → Log. You can view the corresponding error logs, broken down by time periods. The collector runs every hour on every instance of the system.
· To determine the profiles for which the above aggregates are to be created in the workload collector database, choose the following path in the navigation bar: Collector and Performance DB → Workload Collector → TOTAL Collector → Control.
· To check that the corresponding collector is functioning correctly, choose the following path in the navigation bar: Collector and Performance DB → Workload Collector → TOTAL Collector → Log. You can view the corresponding error logs, broken down by time periods. The collector runs system-wide once every three hours.
· To determine how long the different aggregates are to be stored in the workload collector database, choose the following path in the navigation bar: Collector and Performance DB → Performance Database → Workload Collector Database → Reorganization → Control.
Note that, depending on the size of the screen, the important part of the column headers may not be visible, and the time unit for the different aggregates is different. The table consists of the following columns:
No. |
Heading |
Unit |
1 |
Daily Aggregates Retention Period |
Day |
2 |
Weekly Aggregates Retention Period |
Week |
3 |
Monthly Aggregates Retention Period |
Month |
4 |
Daily Aggregates Retention Period (TOTAL) |
Day |
5 |
Weekly Aggregates Retention Period (TOTAL) |
Week |
6 |
Monthly Aggregates Retention Period (TOTAL) |
Month |
· To check that the above reorganization is functioning correctly, choose the following path in the navigation bar: Collector and Performance DB → Performance Database → Workload Collector Database → Reorganization → Log. You can view the corresponding error logs, broken down by time periods. The collector runs system-wide once a day.
Workload Monitor start page