Professional Documents
Culture Documents
1
Objectives
Perform monitoring
2
Monitoring Concepts in General
Solution Manager
End to End
Configuration of the End to End
Performance monitoring
3
End to End - Monitoring
System monitoring
System administrator
Focus on single components SAP APO
Systems have to operate
reliably
4
System Monitoring
SAP R/3
Finance SAP APO SAP BW
System Communication
Database Consistency
Performance …
Alert: Number of open requests
in queue exceeds threshold !!! (activ agents)
5
Process Monitoring
6
Benefits of Process Monitoring
Benchmarking of processes
7
Business Process vs. Technical Process
Business Process
Flow of business actions
Flexible
Mainly controlled by customizing
Wider scope
Longer periods
Contains multiple technical processes
Technical Process
Flow of technical steps
Static with few variations
Mainly controlled by coding
Smaller scope
Shorter periods
8
Definition: Process and Process Steps
Technical
IDoc- TRFC IDoc- Call
Process Out- In- Appli-
Steps bound bound cation
9
Example: Technical Steps in ALE Process
System A System B
Application
ALE API
Create IDoc
asynchronous R
(configurable)
10
Monitoring – Current Situation
APO
11
Example: Current Situation ALE Monitoring
12
ALE Interface Monitoring
Business Process: Sales order information transferred from SAP R/3 System
X48 to SAP R/3 System DSM via ALE interface.
Solution Manager
Red alert :
Interface in
critical state
13
ALE Interface Monitoring - Errors
14
ALE Interface Monitoring - Performance
15
Issues with Existing Tools
16
End to End monitoring (E2E)
Main tasks
Localization of errors and performance
bottlenecks
Single interface for monitoring multiple
technologies (XI, qRFC, ALE, …)
Documentation of process flow CRM R/3
Data Provider for SAP Solution Manager
Benefits
Faster issue resolution
Quality improvements
Easier administration and support APO
=> Reduced TCO
17
End to End monitoring
Goals
Realization of a monitoring infrastructure for technical processes
Support of cross-component and asynchronous processes
Data provider for SAP Solution Manager (Business Process
Monitoring)
Prevailing Conditions
Runs at customer sites with minimal customizing
Web user interface
Supports lower releases
Cooperation with SAP Solution Manager
No (or at most minimal) changes of existing applications
System and application load to be neglected
Unification of existing tools, especially user interface
18
End to End & SAP Solution Manager
Technical
IDoc- TRFC IDoc- Call
Process Out- In- Appli-
Steps bound bound cation
19
End to End monitoring and infrastructure
In the last pages we have gone through the technical concept and
infrastructure of the monitoring.
20
Monitoring Concepts in General
Configuring the End to End Monitoring
End to End
Configuration of the End to End
Performance monitoring
21
Configuring the End to End Monitoring (step 0)
3.On each Adapter Engine that is to be monitored, a destination called pmistore needs
to be defined using the service user for the Runtime Workbench (XIRWBUSER) for
authentication.
For further details, see the Installation Guide – SAP Exchange Infrastructure 3.0,
chapter Creating HTTP Destination pmistore in the SAP J2EE Environment.
Ensure that the service user for the Runtime Workbench (XIRWBUSER) is defined in
each component that is to be included in end-to-end monitoring.
22
End-To-End Monitoring: Configuration II
com.sap.aii.rwb.server.centralmonitoring.r3.ashost
com.sap.aii.rwb.server.centralmonitoring.r3.client
com.sap.aii.rwb.server.centralmonitoring.r3.sysnr
com.sap.aii.rwb.server.centralmonitoring.httpport
ITS plug-in running on the Integration Server and central monitoring server.
To allow navigation from the Runtime Workbench to ABAP monitoring transactions, for
example to RZ20, the ITS plug-in needs to run on the Integration Server and on the central
monitoring server.
Use transaction SICF to check whether the ITS service is activated. You may also
have to set profile parameter em/plobal_area_mb to 256 to allow proper ITS
operation.
23
End-To-End Monitoring: Configuration III
Activation of required SICF services
In addition to ITS, additional SICF services have to be activated on the central monitoring server in order to allow proper
operation of end-to-end monitoring and alerting. You may activate the required services by executing the report
RSXMB_ACTIVATE_ICF_SERVICES.
Alternatively, you can manually activate the services listed below by transaction SICF.
SICF services to be activated
24
End-To-End Monitoring: Configuration IV
The Runtime Workbench as well as other services, i.e. the MDM, the directory cache
refresh, read component and addressing information from the SLD. There are two
complementary ways to provide the SLD with XI-specific content:
Manually
You define business systems by using the Business Landscape UI of the SLD.
Before that, the technical systems need to be described in the SLD.
XI components register themselves with the SLD as soon as they are started up.
This type of self-registration does not require SLD bridges.
25
End-To-End Monitoring: Prerequisites
To active the collection of performance data needed for the monitoring this
parameter has to be set to “1”.
A “0” here will deactivate the collection of data for the End to End monitoring.
26
End-To-End Monitoring: Configuration I
monitoring as a sending or
Select component to be Select the monitoring level receiving party
monitored
After configuration, choose button „Save Configuration“ to activate the End To End
monitoring.
Currently when you save the new settings the old settings are deleted.
Before you can start to user the end to end monitoring you need to configure it. The changes you make will be
noted by the system, this information is the user and the time of the saving of the configuration change.
Therefore you first steps in the system need to be in the Configuration menu. Here you have to make a few
choices.
When you start the configuration screen all components that are registered in the SLD and associated with the
domain of the Integration Server are displayed. This means that any components that are “outside this
domain” can’t be monitored via the End to End monitoring.
Going from left to right:
1) In the display a list of all systems/components that are stored in the SLD are displayed. Here you can
activate the monitoring for a component. You also see on type of component and its name.
2) The monitoring level is a setting as to how much data is gathered for the particular component. There are
three levels (low , medium, high) of data collection. According to the component you choose to monitor the
amount of data can vary. Some component have multiple measuring points, others have two. So the amount
of data depends on how much data you care to have for your analysis.
3) The next choice you have to make is what “path” you wish to monitor. You can choose the incoming,
outgoing or both sorts of messages for that particular component. As not all components are both sending
and also receivers.
4) The last section depends on what components you have chosen to have monitored.
For ALE there are more entries that need to be done, in this case you have the choice to monitor the defined
ports of the ALE communication. If you leave this entry blank then by default the system will monitor all
appropriate dialog on the component.
5) When you save the settings the current status is that the old settings are overwritten and also the data
gathered is also deleted.
6) You also have the option to deactivate the Monitoring, this option lets you stop the End to End monitoring
completely. No data is gathered from the agents nor sent to the central End to End server and aggregated.
27
Possible Errors
28
Monitoring Concepts in General
Configuring the End to End Monitoring
Using the End to End Monitoring
Configuration of the End to End
Performance monitoring
29
End-To-End Monitoring – Initial Screen
To start the performance measuring using he End to End you need to first choose
“End to End Monitoring”.
If you have detailed information on the participating messages you wish to monitor
you can make this choice here. There is neither value help nor wildcard support so
you need the exact spelling of any object you wish to analyze.
If you wish to get a general overview then you can simply press Display. This will
give you the largest detail of information.
30
End-To-End Monitoring: Overview
The next display then has two sections. On the left hand side you can see the
detailed information and on the right you have a overall view of the landscape and
the processes that are sending and receiving the messages.
If you click on the checkbox “display only participating components” the display
gets rid of the non-participating components in the monitoring timeframe.
31
End-To-End Monitoring: left - Overview
This is the Monitoring Status: Here you can see the time of the last
configuration change and by whom it was done. Also you can do
some basic settings here.
The choices here are if you wish to use the process view or the
instance view and if you wish only see the components that
were involved in processing messages in the time frame
chosen.
This is the “quick view” if you had problems with the message flow.
Here we have the tally of how many errors, open, successfully finished
and in total processed messages. In addition to that the time needed
to process is also displayed. (the average, maximum and minimum
values).
Here you can take a closer look at the instance list that is selected via
the statistics or an entry in the graphical overview.
32
End-To-End Monitoring: right - Overview
33
End-To-End Monitoring: Settings - General
34
End-To-End Monitoring: Settings - Details
Notice: To set any values in this panel you need to first disable
the monitoring in the General Screen. By deactivating the
monitoring you are not deleting any performance data gathered.
In this section you can define how long the End to End keeps
the data in the system, before it starts to “compress” it. After
the defined time the system starts the compression of the
data.
Under the header “Resource management” you can choose what level of resources usage you wish to
have for End to End.
If you have defined RFC-server groups then you can assign these RFC-servers to handle the processing.
By default it is to use all RFC-servers in the defined server group.
Also you can define how many dialog work processes are to used for the processing from each server in
the Servergroup. Per default 25% of all work processes (that can be used for RFC calls) from each server
is used.
The settings under the topic “orientation values” are more along the lines of suggested values. There
are currently no defined values. These values are used to define a few internal settings for the End
to End monitoring. Currently the suggestion is to leave the settings where they are.
Number of XI-Messages per day. This value is a estimation of how many messages you expect to have
travel through the XI server.
Average duration of a XI-message: This one is how long a message will probably stay in the XI-pipeline
on average, before it arrives at the target system.
And finally the maximal time that a message can take before the End to End considers it as being open.
As a message can take a long time to get to the target system. This values basically is used to define a
limit of how long End to End will wait for the “closing” of the message. If a message takes longer then this
values it is considered to stay open, and will not the set to finished.
The next set of settings are used for defining the reorganization of End to End data. In principle you
can make any choice you wish here. Keep in mind that the longer the retention period, the more
data is stored in the system. This has of course a direct effect on the End to End monitoring
performance.
35
End-To-End Monitoring: Settings - Error Status
The Short text gives you a short explanation in regards to what happened.
36
Monitoring Concepts in General
Configuring the End to End Monitoring
Using the End to End Monitoring
End to End process viewing
Performance monitoring
37
Process view
Sending systems
XI system
Receiving systems
The process overview is the initial screen of the Process Monitoring and contains
the following important display functions:
Graphical display of the process flow across the components involved (for
example, XI message processing using a sending system, Integration Server and
target system)
Status display and error messages for the process flow
If an error has occurred in the monitored process during a transfer between
components, an error message is displayed, together with red status icons in the
process flow.
Display for how many process instances have run in total, and how many are
open, completed, or have errors. This gives you a quick overview of the process
status.
Selecting the number of open, incorrect or complete process instances displays a
list of the corresponding process instances. This gives you a quick overview of
process instances that have errors, for example.
38
Instance View
d view
aile
det
39
More Details via the magnifying glass
by pressing the +
With the help of the icons in a processing step you can see where there are
issues.
You then can click on the magnifying glass to get a more detailed view of the
processing step. Here you can decide at which step within that component you
need to do further checking.
If you click on a number in a component/step, you get the list of still available
instances contributing to this number.
40
Finding issues with End to End monitoring
1) In the general overview you will notice a red process “box”. This
denotes an error situation within that section. Double click on the
“error flag”. On the left hand side of the display you will see a list
of the “errors” messages.
41
Error tracking I
42
Error tracking II
by choosi
ng the m
essage
detailed
view
By choosing the message with the error a new window is displayed on the right
hand side. it is the instance view for this process.
In this window you can again click on the step in which the error occurred.
43
Error tracking II - Attribute display
44
Error tracking III
In this display you pick the message you wish to analyze further, then click
display.
45
Error tracking IV
This last display then is close to the display generated by the sxmb_moni.
From this point on you can drive your error analysis as you wish.
46
Monitoring Concepts in General
Configuring the End to End Monitoring
Using the End to End Monitoring
End to End process viewing
Performance monitoring
47
How to define Performance Indicators
The problem we have in giving performance values (or suggestions) is that every
customer is different. They differ in:
number of documents
their size and complexity
mapping(s) used and their complexity
how the message is sent from one party to another
specific times needed to send and receive the messages.
and many more items
There is no easy way to suggest a general “good” value for all customers.
There are two ways to find your indicators:
1.During operations you will develop a certain number of values that you can use
as performance indicators.
2.Or your customers define a number of predefined Key Performance Indicators
(KPI).
Whatever form of definition you use, it has to be clear that all parties need to
agree to them and that they are in written form. A regular reporting on these should
also be defined.
48
starting the performance monitoring
To start the Performance monitoring use TA: SXMB_IFR a webpage is generated. There you
choose the Runtime workbench (RWB).
Then choose the item Performance Monitoring in the top row. Then you have choice of what
data source to use the End to End or the IS for the monitoring
49
difference between IS and PMI
The IS view displays the locally stored data for the specific
integration server chosen in the pull down window.
If you choose PMI, the data collected from the PMI server is
displayed according to your settings. You can also choose
other components within the landscape (as long as they are
known to the PMI ) and see their performance data.
50
The Performance Monitor
51
aggregated Overview
Column Meaning
Interval Start/End of Interval Start and end of the interval in question
Number Number of messages processed in this interval
Size Average size of the processed messages
Latency Average processing time in seconds
Rate Number of messages processed each hour
Component Component in which message processing took place
52
aggregated Overview 2
1) timeframe
2) number of messages
This then gives you the processing rate, number of messages per
hour. A unexpected deviation from the standard value in the
processing rate can be seen as a potential issue.
When this happens, compare the average size of the documents and
their volume.
You then may need to implement further steps to resolve the issue. This depends on
your specific scenario as to what you need to do.
53
single Overview
54
aggregated Detail view
You are shown the summed performance sets of all messages that fit the chosen time. According to what
component you are viewing with the End to End you have different timestamps displayed in the display.
If you are accessing the ABAP stack then six timestamps (phases in the XI-pipeline) are available.
PLSRV_RECEIVER_DETERMINATION
PLSRV_INTERFACE_DETERMINATION
PLSRV_RECEIVER_MESSAGE_SPLIT
DB_SPLITTER_QUEUING
PLSRV_MAPPING_REQUEST
PLSRV_OUTBOUND_BINDING
For a component in the End to End two timestamps are available one for entry into the component and one when
leaving the component.
If the component is a adapter then there is one timestamp.
With the given information you can analyze a particular component and its processing speed or is not working
properly.
Column Meaning
Interval Start/End of Interval Start and end of the interval in question
Number Number of messages processed in this interval
Size Average size of the processed messages
Latency Average processing time in seconds
Rate Number of messages processed each hour
<Time stamp> Specific times of processing (with minimum and maximum values, if
selected)
55
aggregated Detail view
56
Summary
Perform monitoring
57