This blog is targeted to BASIS ADMINS
Transaction for workload analysis statistical data changed over time are monitored using transaction code ST03 , now ST03N (from SAP R/3 4.6C) . With SAP Web AS 6.4 the transaction ST03 is available again.
From time to time ST03 and ST03N has seen many changes but later in SAP NW7.0 ST03N has reworked in detail specially processing time is now shown in separate column.
Main Use of ST03N is to get detailed information on performance of any ABAP based SAP system. Workload monitor analyzes the statistical data originally collected by kernel.
You can compare or analyze the performance of a single application server or multiple application server. Using this you start checking from the entire system and finding your way to that one application server and narrowing down to exact issue.
By Default :- You see data of current day as default view , you can change the default view.
Source of the image : sap-perf.ca
Let's discuss the WORKLOAD MONITOR
By Default : Expert mode needs to be used
As you can see in the picture you have all the data in second or millisecond as unit.
The common way to start analyze the issue is looking into workload monitor ,
So you need to check the overview in this flow , Check for Today , Last Minutes Load, Transaction profile and the response time duration, Total workload option , gets data from the whole SAP System.
Now Lets understand the different monitors that are used and their uses,
So we need to check in expert mode , because hey we are the basis guys!
WORKLOAD : Two levels of analysis : Individual (Only one Application server) or Total/ALL (All Application server)
All configured Application servers of the SAP System are listed , even when application server is temporarily shutdown.
The other sub views are Day, Week and Month. So that we can check the performance problem of the past easily.
DETAILED ANALYSIS : Business transaction analysis is STAD , this is the successor of STAT which is no longer supported .
This displays workload statistics for every step executed in ABAP based system
As we can see in the below mentioned image we have 3 display mode :-
1. Sorted by time : Display Single records.
2. Grouped by business transaction : Shows all record , grouped by business transaction.
3. Show Business transaction tots : Display single records grouped by business transaction or jobs.
We have different filter parameter , you can select a user , business transaction or different parameter can be used.
We can use different filter parameters as per our goals.
All records are read but only selective are displayed.
Usually you will be working using the view , Show all stats records, Sorted by time.
Now [*] the option include statistics from memory , The System Analyzes that are not yet written in stats file but which are stored in the stats of buffer. This is must when you need to analyze the recent periods i.e. buffer records is included in this case.
Server Selection allows you to analyze the stats of selected servers of the SAP System.
With AS ABAP 7.02 the transaction STADWD became available. STADWD implements the functionality of STAD using web Dynpro of ABAP.
If you did not receive data from a server that was called as a result of RFC problem or busy AS that increases
Adding here is Another option Last minute's Load again you can either choose a an individual application server or a total summary.
You have Analysis Interval : Read interval [Date, Time]
You can also have data restriction : Filter Parameter [Client, Username and Work process number ]
The granularity of the time interval is determined. This granularity is used in the analysis view Time Profile.
LOAD HISTORY AND DISTRIBUTION :
In this also you will have two views : Individual and Total application server, which can be displayed by date , week or month or any time period i.e. different time interval can be analyzed.
By default only DIA (Dialog task types) are listed but obviously others can also be seen.
You can also check users per instance along with a cumulative view.
BI WORKLOAD :
Business Warehouse capabilities of Netweaver is checked here.
Schedule corresponding collector that gathers respective information about the workload caused by Business Warehouse functions in SAP System.
COLLECTOR AND PERFORMANCE DB :
1. Performance Database
2. Performance Monitor Collector
- access to log files of report RSCOLL00 that collects the data as per scheduling mentioned in table TCOLL
3. Workload Collector
- Customize the workload collector , for example how many records to collect during a run.
4. Statistics records and files : Here you can look up and change SAP System Parameters relevant for collecting statistics data.
Let's understand the ANALYSIS VIEW
We can see different Analysis views present in ST03 that is present in the left pane.
Source - sap-pref.ca
Time Profile : Statistical data of chosen task is displayed segmented according to granularity level and time interval you specified.
source : saptechnicalguru.com
Ranking Lists : You can find the top 40 executed steps executed within the SAP System , according to task types and differentiated between
1. Response time
2. Number of database calls
3. Memory use statistics "Average total memory usage [kB]"
and you can again differentiate between task type.
RFC Profile : Statistics of separate client side and server side RFC activities of the SAP System, you can look up for :-
1. client and server destination
3. called function blocks
User and Settlement Statistics : User and client activity
Front end statistics : Activity statistics by front end
Spool Statistics : for print related loads.
Response time duration : Response time in the SAP Application server are distributed . Used for monitoring SLAs. For instance someone guarantees an average response time of DIA as 3s or better.
Load from External System : For External System Activities.
Web Statistics : related to web reporting caused by Business Intelligence Activities.
DB Connection Statistics : Stats regarding DB access , differentiated by DB user . Major load is caused by connection "Default"
In the next blog we will dive deep in how we analyze the workload.