test

Sun N1 Advanced Architecture for SAP Solutions 5.2.1 User's Guide

Appendix A MDX Queries

N1 AA Analyzer: Server Group Utilization

This section gives a detailed explanation of both MDX queries:

Based on the measured CPU/Memory utilization, the total host utilization within the specified timeframe is displayed and qualified. For example, low, good, or high host utilization. The qualification is based on customizable threshold values %cpulow and %cpuhigh.

Figure A–1 Example of CPU Utilization

Example of CPU Utilization

Looking at the host CPU utilization of netra1 in the specified time frame:

Figure A–2 Example of Memory Utilization

Example of Memory Utilization

Looking at the host memory utilization of netra1 in the specified timeframe:

The parameters mentioned previously are maintained in N1 AA Analyzer customizing. See Basic Setup. The following table summarizes all available parameters within both MDX queries.

Note –

All the parameters given in the table below can be used for server group spanning or server group specific parameterisation.

Table A–1 Server Group Parameters

Constant 

Description 

Default Value 

none.analyzer.constants.%cpulow

Lower threshold for CPU average utilization in percentage. 

If the measured average host utilization is smaller than %cpulow, the qualification is low.

30 

none.analyzer.constants.%memlow

Lower threshold for memory average utilization in percentage. 

If the measured average host utilization is smaller than %memlow, the qualification is low.

30 

none.analyzer.constants.%cpuhigh

Upper threshold for CPU average utilization in percentage. 

If the measured average host utilization is in between %cpulow and %cpuhigh, the qualification is good.

If the measured average host utilization is greater than %cpuhigh, the qualification is high.

75 

none.analyzer.constants.%memhigh

Upper threshold for memory average utilization in percentage. 

If the measured average host utilization is in between %memlow and %memhigh, the qualification is good.

If the measured average host utilization is greater than %memhigh, the qualification is high.

75 

none.analyzer.constants.%cpupeakslower

Lower threshold for CPU peak utilization in percentage. 

If less than %cpupeakslower percent of all measured host utilization values are greater than %cpuhigh, the qualification is some.

5 

none.analyzer.constants.%mempeakslower

Lower threshold for memory peak utilization in percentage.  

If less than %mempeakslower percent of all measured host utilization values are greater than %memhigh, the qualification is some.

5 

none.analyzer.constants.%cpupeaksupper

Upper threshold for CPU peak utilization in percentage. 

If more than %cpupeakslower AND less than %cpupeaksupper percent of all measured host utilization values are greater than %cpuhigh, the qualification is more.

If more than %cpupeaksupper percent of all measured host utilization values are greater than %cpuhigh, the qualification is many.

10 

none.analyzer.constants.%mempeaksupper

Upper threshold for memory peak utilization in percentage.  

If more than %mempeakslower and less than %mempeaksupper percent of all measured host utilization values are greater than %memhigh, the qualification is more.

If more than %mempeaksupper percent of all measured host utilization values are greater than %memhigh, the qualification is many.

10 

N1 AA Analyzer: Create Your Own MDX Queries

This section describes how to create your own MDX queries.

Note –

Created MDX queries are available within the server group only.

These MDX queries are based on the underlying database model. The following picture gives an overview on this model. The fields in cursive letters will be available as measurands for your MDX Queries within the cube.

Figure A–3 Overview of MDX Queries

Overview of MDX Queries

Every measurand taken by the Performance Collector on an N1 AA Client leads to a set of self-contained entries in the Resource Consumption table. The measurand is taken every 15 minutes. Each of these entries describes the resource consumption of exactly one measured application component at the given point in time.

Table A–2 Resource Consumption

ID_PhysicalHost 

ID_Component 

ID_TimeSlice 

ID_CPUTypeGroup 

CPUSeconds 

CPUNormSeconds 

MemoryBytes 

1 

6 

2006-01-01 00:30:00 

1 

0.83 

1.66 

105553920 

1 

4 

2006-01-01 00:30:00 

1 

157.87 

315.74 

177684480 

1 

3 

2006-01-01 00:30:00 

1 

48.86 

97.72 

703455232 

1 

22 

2006-01-01 00:30:00 

1 

10.41 

20.82 

4481024 

1 

5 

2006-01-01 00:30:00 

1 

9.30 

18.60 

0 

1 

1 

2006-01-01 00:30:00 

1 

658.00 

1316.00 

–655630336 

1 

2 

2006-01-01 00:30:00 

1 

14.73 

29.46 

0 

(with ID_Component 1 = Free; 2 = Other; 3 = SAP-D01-CI; 4 = SAP-D01-DB; 5 = user.root; 6 = system; 22 = SAP-E01-D05)

In this example, component #3, which is SAP-D01-CI, has used 48.86 CPU seconds within the 15–minute interval and the memory consumption was 703455232 bytes.

In addition to the MDX keywords, the following N1 AA Analyzer specific placeholders were introduced:

Table A–3 Analyzer Placeholders

Placeholder 

Description 

[$ServerGroup] 

Constant member set of ServerGroup dimension containing a single member, equals the actual server group at runtime.

[$PhysicalHost] 

Variable member set of ServerGroup dimension and PhysicalHost hierarchy, equals the selected set of physical hosts at runtime. Using this placeholder will enable the Physical hosts section in selection screen when executing an analysis.

[$Application] 

Variable member set of Application dimension, equals the selected set of applications at runtime. Using this placeholder will enable the Applications section in selection screen when executing an analysis.

[$Component] 

Variable member set of Application dimension and Component hierarchy, equals the selected set of components at runtime. Using this placeholder will enable the Components section in selection screen when executing an analysis.

[$Consumptions] 

Variable member set as cross product of ServerGroup dimension and PhysicalHost hierarchy and Application dimension and Component hierarchy, equals the selected set of consumptions. Each happened on a dedicated physical host by a dedicated component. Using this placeholder will enable the Dedicated consumptions section in selection screen when executing an analysis.

[$TimeSlice_Minute] 

Variable member set of TimeSlice dimension, equals the selected timeframe at runtime on a minute-based resolution. Using this placeholder will enable the Timeframe section in selection screen when executing an analysis. All time entries are represented in the UTC time zone when displaying analysis results in tables or diagrams.

[$TimeSlice_Hour] 

Same as above, on a hour-based resolution (termed: aggregation). 

[$TimeSlice_Day] 

Same as above, on a day-based resolution (termed: aggregation). 

[$TimeSlice_Week] 

Same as above, on a week-based resolution (termed: aggregation). 

[$TimeSlice_Month] 

Same as above, on a month-based resolution (termed: aggregation). 

[$TimeSlice_Quarter] 

Same as above, on a quarter-based resolution (termed: aggregation). 

[$TimeSlice_Year] 

Same as above, on a hour-based resolution (termed: aggregation). 

[$TimeSlice] 

Same as [$TimeSlice_Minute], but in dispose of raw times written by the Performance Collector daemon.  

Note –

Do not use this placeholder. All time entries in analysis results will be 15 minutes late.

[$constantname]

Constant that is replaced with the value maintained in Analyzer Customizing. See Basic Setup.

The value will be taken from parameter none.analyzer.constants.constantname.

Within a dedicated server group, the value will be overwritten by parameter servergroup.analyzer.constants.constantname only if this parameter is maintained, which is optional.