Enable the Runtime Performance Advisor tool using scripting

WAS v8.5 > Script the application serving environment (wsadmin) > Troubleshooting with scripting
Enable the Runtime Performance Advisor tool using scripting

We can configure the Runtime Performance Advisor (RPA) using wsadmin or the dmgr console. The recommendations display as text in the SystemOut.log file.
IBM recommends using the HPEL log and trace infrastructure. With HPEL, one views logs using the LogViewer command-line tool in PROFILE/bin.
The processing of RPA is governed by various rules and corresponding rule IDs. The following table summarizes the mapping between rule IDs and the advice the RPA would process.

Rule ID Nature of Advice
ServerRule01 No room for new sessions rule
ServerRule02 Live session
ServerRule03 Session read and write size
ServerRule04 Session read and write time
ServerRule05 Servlet engine unbounded rule
ServerRule06 Servlet engine thread pool rule
ServerRule07 ORB unbounded
ServerRule08 ORB pool rule
ServerRule09 DataSource connection pool min and max size rule
ServerRule10 DataSource prepared statement discard rule
ServerRule11 Memory leak detection rule
surgeModeAlert Surge mode rule
poolLowEffAlert Pool low percent efficiency rule
hungConnModeAlert Hung connection alert rule
connLowEffAlert Connection low percent efficiency rule
connErrorAlert Connection error alert rule
LTCSerialReuseViolationAlert LTC serial reuse violation alert rule
LTCNestingAlert LTC nesting rule
LTCConnPerThreadLimitAlert LTC connection per thread limit rule
multiThreadUseViolationAlert Multi-thread use JCA programming Model Violation
xComponentUseViolationAlert Cross component use JCA programming model violation

The Runtime Performance Advisor (RPA) requires the Performance Monitoring Service (PMI) is enabled. It does not require that individual counters be enabled. When a counter needed by the RPA is not enabled, the RPA will enable it automatically.
There is no MBean/object available for wsadmin to create a RPA configuration. We can use wsadmin to change the settings and make them effective at runtime. These changes will not be persisted. The changes remain until you stop the server. Since the RPA is disabled once you stop the server, you may want to disable the PMI Service or the counters that were enabled while it was active. We can enable the following counters using the Runtime Performance Advisor:
ThreadPools (module)
Web Container (module)
Pool Size
Active Threads
Object Request Broker (module)
Pool Size
Active Threads
JDBC Connection Pools (module)
Pool Size
Percent used Prepared Statement Discards
Servlet Session Manager (module)
External Read Size
External Write Size
External Read Time
External Write Time
No Room For New Session System Data (module)
CPU Utilization
Free Memory

The following provides an explanation for some of the settings that we can use:

Calculation interval PMI data This setting is taken over an interval of time and averaged to provide advice. The calculation interval specifies the length of the time over which data is taken for this advice. Details within the advice messages will appear as averages over this interval.
Maximum warning sequence Number of consecutive warnings issued before the threshold is relaxed. For example, if the maximum warning sequence is set to 3, then the advisor only sends three warnings to indicate the prepared statement cache is overflowing. After that, a new alert is only issued if the rate of discards exceeds the new threshold setting.
Number of processors Number of processors on the server. It is critical in order to ensure accurate advice for the specific configuration of the system.

Setup the Runtime Performance Advisor (RPA), for example:

Jacl:
set perf [$AdminControl queryNames mbeanIdentifier=ServerRuleDriverMBean2,process=server1,*]
set enabledVal [java::new java.lang.Boolean true]
set attr [java::new javax.management.Attribute enabled $enabledVal]
set perfObject [$AdminControl makeObjectName $perf]
set ObjectArray [java::new {java.lang.Object[]} 1]
set sigArray [java::new {java.lang.String[]} 1]
$ObjectArray set 0 $attr
$sigArray set 0 "javax.management.Attribute"
$AdminControl invoke_jmx $perfObject setRPAAttribute $ObjectArray $sigArray
$AdminConfig save

After completing the previous steps, start the server and monitor RPA.

Related concepts:

Performance and Diagnostic Advisor

Related
Use the wsadmin scripting AdminControl object for scripted administration
Troubleshoot applications with HPEL
Use the Performance and Diagnostic Advisor
Start the wsadmin scripting client using wsadmin.sh
Commands for the AdminControl object using wsadmin.sh

+
Search Tips | Advanced Search

Rule ID	Nature of Advice
ServerRule01	No room for new sessions rule
ServerRule02	Live session
ServerRule03	Session read and write size
ServerRule04	Session read and write time
ServerRule05	Servlet engine unbounded rule
ServerRule06	Servlet engine thread pool rule
ServerRule07	ORB unbounded
ServerRule08	ORB pool rule
ServerRule09	DataSource connection pool min and max size rule
ServerRule10	DataSource prepared statement discard rule
ServerRule11	Memory leak detection rule
surgeModeAlert	Surge mode rule
poolLowEffAlert	Pool low percent efficiency rule
hungConnModeAlert	Hung connection alert rule
connLowEffAlert	Connection low percent efficiency rule
connErrorAlert	Connection error alert rule
LTCSerialReuseViolationAlert	LTC serial reuse violation alert rule
LTCNestingAlert	LTC nesting rule
LTCConnPerThreadLimitAlert	LTC connection per thread limit rule
multiThreadUseViolationAlert	Multi-thread use JCA programming Model Violation
xComponentUseViolationAlert	Cross component use JCA programming model violation

Calculation interval PMI data	This setting is taken over an interval of time and averaged to provide advice. The calculation interval specifies the length of the time over which data is taken for this advice. Details within the advice messages will appear as averages over this interval.
Maximum warning sequence	Number of consecutive warnings issued before the threshold is relaxed. For example, if the maximum warning sequence is set to 3, then the advisor only sends three warnings to indicate the prepared statement cache is overflowing. After that, a new alert is only issued if the rate of discards exceeds the new threshold setting.
Number of processors	Number of processors on the server. It is critical in order to ensure accurate advice for the specific configuration of the system.