Information on Performance Monitoring
The Performance in the SAP is calculated based on the average dailog response time. The average dailog response time is calculated once a user request is received by the dispatcher and the dispatcher has send the user request.
SAP recommends threat if the average dailog response time of a SAP system is around 1 sec/1000 milli sec. Then it is set be that the system is having a good performance. Average Dailog response time is calculated based on the below formula.
Average Dailog response time = Total number Dailog step/Total amount of taken in Sec.
The below are factors which influences the system performance.
1. The Network Bandwidth.
2. SAP Resources.
3. Database Recourses.
Average Dailog Response time = wait time + Roll in time + load & Generation + buffer Waite time + Enque time +Database Acese time + Roll out time.
The performance timing of an SAP system basically divided into 2 types.
1. Basis Tuning: Which is done by the Basis Adm in the areas such a as Network.
2. Application Hardware Resources.
3. Database Server.
Application / ABAP Tuning:
This is performed by the abapers to fine tune the SAP Abap programs, reusability of the function modules, classes and variables.
ST04, ST06 and STO3N.
RFC time and CPIC time is the amount of time taken by the user to reach to the application server. Then the RFC time is not part of the average Dailog response time. If it is taking more than 150 M sec. Then we need to consider checking the network bandwidth more in details. ST06.
Wait time is the amount of time waited in the Queue before has been assigned to a work process SAP recommends that the wait time should be use the 10% of the average dailog response time is 150 M sec.
If work process is waiting time we need to increase the work process.
Roll in is the amount of time taken by work process to copy the user master data into the work process internal memory.
SAP recommends to theat 40 -50 M. sec.
The Roll in time is high
Database is busy
Connection between application server and database.
Database is not good.
Application Server: 70%
Database Server: 30%
LOAD AND GENERATOR:
Load and Generator is the amount of the time taken by work process to copy and generate or to load and generate abap code and screens for the User request. The load and generator time is high the problem is the buffer sizes is too small (TP program or buffers and CUA buffers)
32 bit Non – unicode programme Buffer – 150 M.sec to increase 260 M.sec.
62 bit Unicode programme Buffer – 300 M.sec to increase 600 M.sec.
BUFFER WAIT TIME:
Buffer wait time amount of time taken copying user data from the buffer it is data from the buffer it is data from the buffer it is assumed in 0 M.sec.
If the buffer is taking time too much of work load.
Enque time taken by the enque wait time to lock object it should.
If the Enque time is high there are lock by user long time.
SM12 – sap lock.
DATABASE ACCESS TIME:
Database access time is the amount of time taken by work process to copy the data from the database. If Database time is high, that means there is a hardware bottle neck on the database server.
ROLL OUT TIME:
Roll out time is the amount of time taken to write the user data back to the Roll Buffer. If high the roll buffer size is small. The transaction codes analyse the load in the local eneque. STO3N.
ST01 – System Trace.
ST02 – Buffer Analysis.
ST03 – Work Load Analysis.
ST04 – Database Monitory.
ST05 – SQL Trace.
ST06 – Operating System Performance monitor.
The System performance data will be initially written to the Operating System level file path: User/sap/sid/ins/data.
These files are called as statistical files which are copied into the database by the background job.
In order to see the system performance data at SAP level operation modes to be configured for the instance and two background jobs to be scheduled in the server.
1. SAP_COLLECTOR_FOR_PERFMONITOR – RSCOLLOO – Collection of data.
2. SAP_COLLECTOR_NON_R3_STAT – RSN3_STAT_COLLECTOR Statistics.
1. Work Load.
2. Load History
· Select Instance
Transaction: ST02 – Buffer Analysis
1. It should be 95%
2. CUA 10000% – SWAP
Hit Ratio – 95%
Swap Size – 10000/per daily
ST04 – Database Performance
ST05 – Operating System Monitor
ST06 – SQL TRACE
SR13 – To configure help