Check Table Size and Structure of ASR Aggregates

Procedure

If the volume of ABAP statistics aggregates saved in the BW of your SAP Solution Manager system (see Troubleshooting in BW-Based Reporting of ABAP Statistics), you must reduce it. First examine the tabeens to see which profiles, granularities and dimensions require the most space in the database. The terms used have the following meanings:

	Meaning
Profile	The ABAP statstics data (ASR) aggregates are stored in profiles. Profiles represent an aspect of system load, comparable with the analysis views in the System Load Monitor (transaction ST03). There are 14 profiles.
Granularity	The granularity is the data resolution time unit. ASRs have the following granularitities: Day, week and month. Their technical names (as displayed in the program SMBI_ASR_BW_ANALYZE) are: DAY DAY (DD) WEEK DAY (WD) MONTH DAY (MD) Daily, weekly and monthly aggregates in hourly resolution DAY HOUR (DH) WEEK HOUR (WH) MONTH HOUR (MH) Note For more information about aggregation, see Aggregation. End of the note.
Dimension	There are several dimensions per profile. A dimension is a group of one or more attributes of a topic. The program SMBI_ASR_BW_ANALYZE usually outputs a key figure in a column, for each dimension, to tell you the relative memory required for the information. An example of a dimension is the RFC destination called for RFC aggregates, which is saved with the aggregates by default. You can analyze by RFC destinations. If you do not need to do so, you can discard this dimension to reduce the RFC profile data volume. Caution If you use an SAP HANA database, no dimension tables are used when you save the profiles, so the program SMBI_ASR_BW_ANALYZE shows no dimensions. Instead, the number of different values of the most important attributes is shown, which is of similar significance for the memory requirement; the more different values there are for an attribute, the more memory required to retain this attribute in aggregation. End of the caution.

Meaning

Profile

The ABAP statstics data (ASR) aggregates are stored in profiles. Profiles represent an aspect of system load, comparable with the analysis views in the System Load Monitor (transaction ST03). There are 14 profiles.

Granularity

The granularity is the data resolution time unit. ASRs have the following granularitities:

Day, week and month. Their technical names (as displayed in the program SMBI_ASR_BW_ANALYZE) are:
- DAY DAY (DD)
- WEEK DAY (WD)
- MONTH DAY (MD)
Daily, weekly and monthly aggregates in hourly resolution
- DAY HOUR (DH)
- WEEK HOUR (WH)
- MONTH HOUR (MH)

Note Note

For more information about aggregation, see Aggregation.

End of the note.

Dimension

There are several dimensions per profile. A dimension is a group of one or more attributes of a topic. The program SMBI_ASR_BW_ANALYZE usually outputs a key figure in a column, for each dimension, to tell you the relative memory required for the information.

An example of a dimension is the RFC destination called for RFC aggregates, which is saved with the aggregates by default. You can analyze by RFC destinations. If you do not need to do so, you can discard this dimension to reduce the RFC profile data volume.

Caution Caution

If you use an SAP HANA database, no dimension tables are used when you save the profiles, so the program SMBI_ASR_BW_ANALYZE shows no dimensions. Instead, the number of different values of the most important attributes is shown, which is of similar significance for the memory requirement; the more different values there are for an attribute, the more memory required to retain this attribute in aggregation.

End of the caution.

Perform this analysis with the program SMBI_ASR_BW_ANALYZE. It must run in the system and client of the associated BW. Proceed as follows:

Call transaction SE38, enter SMBI_ASR_BW_ANALYZE in the Program field, and choose (Execute).
Note
You can also call the program from the SAP Solution Manager system using the Monitoring and Alerting Infrastructure analysis tools. Proceed as follows:
1. Call transaction MAI_TOOLS.
2. Switch to expert mode by choosing Activate Expert Mode.
3. On the Analysis tab page, choose Analyze ASR Data in BW .
End of the note.
The number of data records is displayed by ASR profile. The PROFILE column contains the profile name.
Check a profile by double-clicking on its row. Most significant sind profiles which use the most data records (RECORDS_TOTAL).
The next screen shows the memory usage (RECS_TOTAL) for the various ganularities, for the selected profile. You can see which granularity has a particularly high memory usage, and should be examined further.
A key figure is output for each dimension. They are displayed in the table in the columns between DUMMY (which is only for the optical structuring of the columns), and the names of the DSOs (DSO_CUBE_NAMES). The greater this value, the more memory required by the granularity to save the data resolved by this dimension (e.g. user names in the dimension DIM_USER).

Note
If you use an SAP HANA database, the number of different values for the most important attributes is displayed, instead of the key figure per dimension. Which attributes these are is specified for each ASR profile, per dimension, in the configuration settings table, under Info Objects Affected. Here, too, the greater this value, the more memory is required to store information according to this attribute.

End of the note.

Choose a granularity to analyze, by double-click.
The next screen shows the memory usage by data from the various managed systems for the selected profile and granularity. You can drill down to system level by double-click. The structure of the data from different systems with similar usage profiles will be similar.
Choose a system to analyze its data, by double-click. The most important elements for analysis are again those with the most data records.
The next screen shows the relative memory usage for the various dimensions or attributes of the selected profile, for the selected granularity, for the selected system. The greater the value of this key indicator, the greater the amount of memory used to save the data by this dimension. This key indicator should not be much greater than 10.000.

With this program, you can determine for which profiles, granularities, managed systems and dimensions the memory usage by ASRs in BW is particularly high, and likely to cause system performance problems. You have determined which configuration settings you can change to reduce memory usage, and thus system load.

How to change these settings is described in Change ASR Aggregate Extraction Settings. You can activate and deactivate data storage by dimension, per managed system.