Monitoring Tools for ERP Applications

I l @ ve RuBoard

ERP applications are used for supply-chain management within an organization. The top ERP applications are SAP R/3, Baan, PeopleSoft, and Oracle applications. This section describes the monitoring tools that have been customized for ERP applications, emphasizing SAP R/3, because it is the leader in this market.

Some ERP application vendors provide a monitoring capability, such as SAP's Computing Center Management System (CCMS) management tool and the EarlyWatch program, which periodically check SAP performance. However, other vendors provide more complete monitoring capabilities. This section focuses on Envive Corporation's monitoring capabilities for SAP, and three monitoring vendors , BMC Corporation, Hewlett-Packard, and Compuware Corporation, who provide customized monitoring products for multiple applications. Emphasis is placed on HP-UX solutions, because HP 9000s are the leading Unix platform for ERP applications from PeopleSoft, Baan, SAP, and many other key applications.

As with other ERP applications, the availability of SAP R/3 depends not only on the application, but also on the underlying database, network, and OS. SAP R/3 is a multi-tiered application with an application server, print servers, and a database server. Properly monitoring SAP R/3 requires the monitoring of all of these components . Products such as SMART Plug-Ins and BMC PATROL are advantageous because of the breadth of monitors that they have available.

Envive

Envive provides several tools for monitoring SAP R/3:

  • Inspector for R/3: Manages SAP R/3 performance and availability, provides root-cause analysis and predictive analysis, and recommends recovery actions. The product analyzes 40 to 50 different categories of performance problems and collects and stores its information in a database on a dedicated NT server.

  • Service-Level Manager for R/3: Reports global SAP R/3 usage, performance, and availability information. Reports can be generated comparing actual service levels to service-level agreements that have been defined. Predefined reports are available that can be scheduled to be generated at regular intervals. Response time, transaction rate, and total bytes transferred are examples of the types of statistics that can be included on these reports .

  • StopWatch: Provides end-user response time information for SAP R/3. For every SAP R/3 transaction, StopWatch collects the total round-trip response time, which includes network time, SAP R/3 time, and client idle time. Reports are provided by user , transaction, or network segment. StopWatch is installed on the same LAN segment as the SAP R/3 application server so that it can capture LAN traffic and correlate logical transaction statistics.

Envive claims to be able to detect 1500 different SAP R/3 symptoms, including deadlocks and a hung application. The applications have been certified with HP's MC/ServiceGuard.

SMART Plug-Ins

IT/O's ability to detect availability problems for an application or server was discussed earlier in the chapter. IT/O is meant to provide a consistent interface for managing systems, applications, and networks. The goal of HP's SMART Plug-Ins (SPIs) is to extend the monitoring capabilities of IT/O by providing preconfigured monitors for applications and databases. SMART Plug-Ins solutions are meant to be used with IT/O with minimal or no configuration. Corrective actions are also predefined for many SPI events. Tools such as BMC PATROL have similar capabilities, but are not as tightly integrated with IT/O.

IT/O can make it easy to deploy monitors on multiple systems. Monitoring templates are distributed from a central IT/O management station. The monitors can also be integrated with MeasureWare and PerfView, so that you can correlate performance data with system and database performance data.

SMART Plug-Ins are available for both SAP R/3 and Baan ERP applications. The SAP SPI can monitor the following conditions:

  • Errors in log and trace files

  • Status of the processes in R/3 instance

  • Dialog task frequency, response time, and wait time

  • Update task frequency, response time, and wait time

  • Batch task frequency, response time, and wait time

  • Spool task frequency, response time, and wait time

The SAP SPI can also send batch job alerts for jobs that were aborted, ran longer than expected, or failed to start. It can also send the following CCMS alerts:

  • ABAP/4 (the programming language for SAP R/3) database events

  • R/3 instance buffer problems

  • Tracing information

  • Enqueue server messages

  • R/3 general messages

  • Rolling and paging activities

  • Internal R/3 database events

  • Configuration status information

  • Detailed information for system log file messages

The SAP R/3 SPI is tightly integrated with the SAP CCMS administrative tool. CCMS can be launched in response to a reported problem. Other management tools can also be launched from the IT/O console. The SAP server is supported on both HP-UX and Windows NT systems. The product can be used with the HP OpenView Manager for SAP R/3, which can generate reports on application servers and database uptime.

SPIs for Baan IV can monitor the status of Baan applications. Along with monitoring the status of each Baan instance, SPIs for Baan IV can monitor log files, license usage, and shared memory usage. Corrective actions can be taken in response to alerts, such as restarting a Baan instance or printer daemon. Reports can be generated on the Baan service's availability or license usage. The Baan managed node must be an HP-UX server. Baan has specially certified this SPI as a product that is well-integrated with the Baan application.

In addition to the ERP applications, plug-ins are available for Oracle, Informix, and NT Internet server. Hewlett-Packard plans to increase the number of SPIs available.

BMC PATROL Knowledge Modules

BMC Software provides monitoring capabilities through its PATROL software suite. PATROL is a system, application, and event management suite for system and database administrators. PATROL provides the basic framework for defining thresholds, sending and translating events, and so forth. Optional products called Knowledge Modules (KMs) contain the capability to monitor specific components. For example, BMC PATROL includes KMs for Unix, SAP R/3, Oracle, Informix, PeopleSoft, and other applications. In fact, more than 40 KMs are available from BMC for use with PATROL.

PATROL consists of a console, intelligent agents , and KMs. The console provides a centralized graphical display in which icons represent system components or other monitored components. Icons change color to correspond to status. The PATROL console also provides an event browser, where you can filter, correlate, sort , and escalate events. Alarms can be configured so that events are sent to the console, indicated graphically, and shown in the event browser.

Intelligent agents provide the ability to discover the system, database, and application components in the enterprise. The agents reside on each server. On an ongoing basis, the agents look for problems, and when they locate one, they either take preconfigured actions or send notification so that recovery can be done manually.

The KMs contain the expertise used by PATROL to know what to monitor and how to react when problems occur. KMs are used to monitor a set of "parameters," which can include a description of the monitored attribute, the polling interval, the method for measuring the attribute, and a threshold for abnormal values. The KMs provide rules to detect events and perform corrective actions. Events are sent to an operator console when an error or warning condition occurs.

BMC provides tools in its Unix KM that can help in monitoring applications. For example, the ACTIVEPROCESS tool monitors active processes. The PROCESS tool displays statistics on active and zombie processes. The USER tool monitors user sessions and user processes. Each of these tools appears with a special icon in the PATROL GUI. In addition to its generic application-monitoring capability, BMC provides specialized KMs for the key ERP applications. Application KMs include SAP R/3, Baan, and PeopleSoft.

PATROL's SAP KM replaces the monitoring aspect of SAP's CCMS management tool. The KM displays CCMS ABAP alerts, buffer alerts, dispatcher queue alerts, enqueue alerts, operating system alerts, and other CCMS alert messages.

Monitoring the R/3 system and application servers is important for SAP. The gateway service is also important. The System, Database, and SAP R/3 Application KMs should be used together to monitor the SAP application, because each KM monitors important metrics. After configuration, BMC discovers the R/3 instances and allows the operator to view an SAP instance and its associated databases and systems. BMC has worked closely with SAP AG to get access to additional SAP R/3 information through a set of ABAP functions. BMC PATROL also provides administrative actions for SAP R/3.

With the SAP R/3 KM installed, the PATROL console displays custom icons for each R/3 system, R/3 instances, CCMS alerts information, printing saplpd processes, R/3 users, tablespaces, and performance information. Database and system icons are also provided.

SAP failures are often due to running out of free space. The System KM can check the filesystem space and available memory. The System KM also supplies important performance metrics, such as network I/O and disk I/O rates.

The key SAP R/3 attributes being monitored include the following:

  • Status of the R/3 instance

  • Historical trend of user activity for an R/3 instance

  • Number of active batch processes

  • Number of active dialog processes

  • Number of active update processes

  • Number of active spool processes

  • Number of active enqueue processes

  • Users currently logged in to the R/3 instance

  • Printer status and queue length

  • Buffer performance information

  • R/3 buffer storage utilization

  • Work-process frequency, response times, and average wait times

  • System log information

  • Dispatcher queue size

  • Transaction times

  • Enqueue errors

SAP R/3 is built on an RDBMS, so it's also important to monitor the database. The key database metrics provided are:

  • Tablespace free space

  • Segments allocated to target database tablespace

  • Free space deficit

  • Archive log free space

  • Cache hit ratio

  • Dictionary cache hit ratio

  • Open cursors

  • Transaction rate

  • Processes used

  • Active transactions

  • Block change rate

The available database metrics, such as transaction rates and cache hit ratios, are described in detail in Chapter 8, as is system information, such as filesystem capacity, CPU utilization, and memory paging information. Important system metrics are described in Chapter 4.

PATROL presents various application "views" to help you monitor each R/3 function. For example, a printer view shows all the printers for a target R/3 system. Status is indicated with icons that are different colors. The first view shows all the target R/3 systems that have been discovered . From this view, you can determine the type of database being used and the number of users currently accessing the R/3 system. Other views exist for printers, memory, CPU, swap, database information, R/3 instance and user information, CCMS alerts, and performance. In each view, you can click icons to obtain additional information.

BMC has been certified to run in an MC/ServiceGuard environment, but issues of inconsistency have arisen. For example, PATROL may report an SAP process failure to its BMC Console, while MC/ServiceGuard may have already restarted the process or moved it to another system.

The PATROL KM for SAP R/3 is supported on Sun Solaris, IBM AIX, and HP-UX, among other platforms.

EcoSYSTEMS

Compuware Corporation provides the EcoSYSTEMS suite of management tools, which includes EcoTOOLS for system and network management, EcoSCOPE for monitoring application performance, and EcoSNAP for managing application failures. The products are supported on both NT and Unix platforms, with HP-UX, IBM AIX, SunOS, and Sun Solaris among the supported Unix platforms.

EcoSCOPE can automatically discover which of hundreds of predefined applications are running on the network, and the user can add application definitions to the EcoSCOPE database. EcoSCOPE can then monitor the number of users accessing an application, the application's response time, byte counts for each application, and the traffic load per application.

A key feature of EcoSCOPE is its ability to monitor application response times. No application changes are needed to get the response time information. EcoSCOPE provides application- or session-level response times, including Oracle transaction information. The response times for individual verbs in an Oracle or Sybase SQL statement can be monitored, to help diag nose long database response times. Minimum, maximum, standard deviation, and mean response time information can be displayed.

Compuware's EcoSCOPE can analyze application traffic at the protocol level. You can see network traffic by application and by protocol. You can recognize multicast and broadcast packets and determine the application sending them. The traffic loads between two nodes can be displayed per application.

EcoTOOLS is used for network and system management. It monitors resource utilization and system availability, and provides both historic and real-time analyses. Alerts can be sent via console messages, pages, or e-mail. Corrective actions are provided, such as restarting a server, re-establishing network connections, or restarting a key application or database. EcoTOOLS can also provide a history of the events leading up to a failure.

EcoTOOLS centrally collects data from the software agents that reside on each server. Support is provided for more than 600 agents. Alarms can be viewed from the master console. Events and data are stored in centralized or distributed repositories. From the central console, the user can also view hardware and software inventory. A GUI enables users to choose new resources to monitor and set thresholds.

Monitoring information is provided at the application, database, network, and OS levels. The system information being monitored includes CPU utilization, disk utilization, and memory usage (such as available swap space). EcoTOOLS can identify the top memory consumers, runaway processes, or shortages in key system resources, such as process table entries, semaphores, and inode entries. Hung printers or peripheral devices can also be detected . System security is provided by verifying that user passwords are installed, checking for unauthorized attempts to access resources, and checking for modifications to system files and directories.

EcoTOOLS provides specific monitoring analysis for the following:

  • SAP R/3

  • PeopleSoft

  • Oracle applications

SAP R/3 services are monitored for their memory performance, CPU performance, and availability. In an HP-UX Oracle environment, additional monitoring is available. EcoTOOLS can monitor the number of SAP users, the application response time and throughput, the buffer usage and configuration status, server workload, and CCMS alerts.

For PeopleSoft, EcoTOOLS can monitor the health of the database. Orphan processes that are created when PeopleSoft application users shut down their systems can be detected and killed . EcoTOOLS can also monitor the dispatch queues of Tuxedo, which is commonly used with PeopleSoft.

The available Oracle tablespace can be monitored. Also, when near limits, additional space can be added. EcoTOOLS monitors the Oracle alert log and concurrent request logs by looking for user-specified strings and sending alerts when they are found. Key Oracle back ground processes are monitored as well. Performance information, such as the number of pending jobs and long-running jobs, can be monitored for the Oracle Concurrent Manager.

Reports can be generated on resource utilization, traffic load, and response time. Users can also define their own reports. Reports on resource usage patterns, for example, can be used for capacity planning. Service-level agreements can be implemented by allowing application and database performance requirements to be defined, and then they can be monitored to ensure that the requirements are met. Exception reports can also be generated.

EcoTOOLS has done specific integration work with the Tivoli and CA Unicenter management frameworks. The EcoTOOLS Plus Module for Tivoli TME 10 forwards events to the Tivoli Enterprise Console. Tivoli templates and filters are incorporated, which make it easier to run the EcoTOOLS GUI and distribute agent software from the Tivoli Enterprise Console. Similar integration has been done for CA Unicenter.

I l @ ve RuBoard


UNIX Fault Management. A Guide for System Administrators
UNIX Fault Management: A Guide for System Administrators
ISBN: 013026525X
EAN: 2147483647
Year: 1999
Pages: 90

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net