19.4    Monitoring

Because SAP PO is mostly used for business-critical processes, it’s important to monitor and keep track of what is happening to the SAP PO system. Monitoring of your SAP PO system can be achieved in several ways. The different possibilities to monitor both SAP BPM and AEX were outlined in Chapter 15.

As you might recall, it’s possible to monitor the message processing between the AEX and SAP BPM by using the following monitors:

These monitors can be accessed via http://<hostname>:<port>/pimon.

To monitor and manage SAP BPM processes, you have to access SAP NetWeaver Administrator via http://<hostname>:<port>/nwa and then choose Operations • Processes and Tasks. From there, you can perform the following tasks:

The monitoring options summarized previously were explored in more detail in Chapter 15. The remaining sections will explore additional monitoring possibilities that we haven’t yet discussed.

19.4.1    Runtime Workbench

The Runtime Workbench (RWB) is a central entry point to access several functions and overviews in the SAP PO system. It can be very useful for quick insight into the high-level status of certain components (see Figure 19.10).

Status of Components within the Runtime Workbench

Figure 19.10    Status of Components within the Runtime Workbench

It can also be used to view the status of the System Landscape Directory (SLD) registration or to execute SLD registration for components when required. Because the SLD registration is crucial for the SAP PO instance to function correctly, this is one of the first places to check in case of issues (see Figure 19.11).

Runtime Workbench SLD Registration

Figure 19.11    Runtime Workbench SLD Registration

Because each component of the SAP PO system needs to be registered in the SLD, all items should have a green status. This screen can also be used when something in the host or port configuration has been changed, and the registration needs to be executed again.

You can access the RWB via http://<hostname>:<port>/rwb.

19.4.2    Wily Enterprise Manager

Wily Enterprise Manager (Wiley EM) is delivered as a standard component within the SAP support contract and is free of charge. Within Wily EM, there are many standard dashboards that can help you analyze the current state of the Java stack. It also caters for historical analysis by providing historical data up to 30 days in the past.

It’s equipped with a console view that shows the high-level status of the system. For a deep-dive analysis, it provides a tool known as the Investigator. Wily EM also provides detailed insight for every node of the Java instance. The next sections will provide more insight into some of its features.

Performance Analysis Using Wily Enterprise Manager

Wily EM can be used to narrow down performance and stability issues within the Java stack, and it has some standard dashboards to monitor the most important key performance indicators (KPIs) of your Java stack. It also provides insight into the queues of the AEX and communication channels.

Besides being used for troubleshooting issues, Wily EM can (and should) be used to perform regular health checks on the system.

Monitor System and Application Thread Utilization

All activities within the Java stack are performed via threads. When the SAP PO system starts and thereafter on request, an initial and minimum thread count is allocated. Additional threads can be progressively allocated until the maximum usage has been reached. If the maximum number of threads becomes allocated and in use, then no additional processing can take place, and the system will basically come to a standstill.

In many cases, it’s sufficient to stick to the default number of system and application threads, but it’s important to regularly perform health checks on the system to check if the number of allocated threads remains stable or sufficient.

If you find that the thread utilization on the system is high, then further analysis should be performed to identify the root cause. Several reasons can justify a high number of threads being used by the system. Applications that weren’t properly closed, hanging sessions, a bug in the software, and so on can be among the possible reasons.

A system running out of its capacity can also justify such an issue. In that case, one of the possibilities to resolve the issue will be to configure an additional server node or install an additional application server.

Monitor Java Heap Utilization

All processing within the Java server node takes place in the JVM. The heap memory forms an important part of the JVM. When a process is started, it claims memory in the heap space to process requests. The garbage collection process is responsible for checking the heap memory at regular time intervals to verify whether the allocated memory segments (in the heap space) are still allocated to an active process. If that isn’t the case, then they will be moved to another space segment within the heap or be cleared.

19.4.3    SAP Management Console

The SAP Management Console (SAP MC) (see Figure 19.12) provides a common framework for centralized system management. SAP MC can be accessed via the following URL, which will launch a Java applet from the browser: http://<hostname>:5<instance_no>13.

SAP MC is responsible for the following tasks:

SAP MC

Figure 19.12    SAP MC

19.4.4    SAP Solution Manager Monitoring

Within SAP Solution Manager 7.1, there is specific monitoring functionality implemented for the monitoring of the SAP PO usage type. The main advantage of using the SAP Solution Manager monitoring tool is its ability to automate monitoring. It requires configuring thresholds on different system parameters in SAP Solution Manager. If those thresholds are reached, then an alert is triggered and sent to the administrator, so there’s no need for the system to be continuously or actively monitored by a human. The administrator’s interaction is needed when an alert event is received.

If you receive too many alert events on a particular interface or the technical system, then this can be an indication that the monitoring thresholds have been set incorrectly or that the SAP PO system needs tuning. The process of monitoring is a cyclic process, which means that it’s never finished. The monitoring configuration needs constant attention and needs to be adjusted according to the system behavior and business requirements.

Technical Monitoring provides a set of metrics on the technical components of your system landscape. This feature provides an overview of the overall system status. The most important items that can be monitored from this overview include the following:

The preceding metrics only represent a subset of all available metrics. These metrics are delivered via standard templates and can easily be adjusted to your own needs. You can add or remove metrics or even define your own metrics and thresholds.

The components that are available within the SAP PO monitoring part of SAP Solution Manager are as follows:

19.4.5    Tracing

Trace information is generally used by developers or system administrators for debugging purposes to diagnose common problems with software. This section will explore some aspects of collecting traces on SAP PO.

Client-Side HTTP Tracing

HTTP and SOAP communication channels can easily be traced without the need for an external tracing tool. It’s possible to enable tracing per receiver channel by setting the XISOAPAdapterBean parameter in the Module Configuration screen in the communication channel, as shown in Figure 19.13.

Module Configuration for HTTP Tracing

Figure 19.13    Module Configuration for HTTP Tracing

The following parameter values can be used for XISOAPAdapterBean:

XPI Inspector

XPI Inspector is a diagnostics tool developed by SAP that can be accessed via a web interface. The tool can collect different types of information from the system, including configuration and traces, and it can perform configuration checks. To be more precise, the following types of information can be collected using XPI Inspector:

Consult SAP help to find the link to download the XPI Inspector software. After downloading the software, it can be deployed onto the Java stack via SAP Software Deployment Manager (SDM), SAP NetWeaver Developer Studio, Telnet, or Software Update Manager (SUM), depending on the release of your SAP PO system. Once deployed, the XPI Inspector can be accessed via http://<hostname>:<port>/xpi_inspector.

19.4.6    JVMMON

JVMMON is a monitoring tool that’s delivered as part of the SAP JVM. The tool provides a wide range of monitoring tasks. The JVMMON application can be found in the JVM directory and is delivered with a command-line tool and a separate tool with a graphical user interface.

The JVMMON application can be attached to a running SAP JVM and provides comprehensive monitoring information or controls various aspects of the JVM configuration. After selecting a running JVM, the following information about its current state is presented:

Furthermore, the following actions can be performed on the JVM:

When the system is hanging or very poorly performing, and past thread dumps are no longer available in the system, you might need to generate dumps of the current situation. These thread dumps are often required by SAP support to help them troubleshoot customer issues.

When a very high amount of memory is allocated within a server node of the Java stack, the server node might generate a heap dump: the complete contents of the memory will be written to a dump file on the operating system level. In such a case, the server process will generally be restarted automatically by the startup and control framework. Heap dumps can also be performed manually using the JVMMON tool.

To analyze the contents of a heap dump, the memory analyzer tool can be used. The Java heap analyzer can be found within SAP NetWeaver Developer Studio as an Eclipse plugin.