How to monitor interactive response times (PEGA0001 alert)

Benefits

C-2359 C-2401 DORID 4.2SP6 Through a prconfig.xml file entry or Dynamic System Setting, you can cause the system to add a line to the alert log when server response time to an interactive browser user exceeds a threshold in milliseconds. This may help isolate the circumstances or user requests that lead to unsatisfactory responses.

When the limit is exceeded, the alert log contains a message of this form:

PEGA0001*Browser interaction time has exceeded a threshold setting.

with additional details.

This measures PRPC server response time only, from the time the HTTP message is received from a user to the time a response HTTP message is sent. The response time experienced by the user includes additional delays, including network delays and workstation time. Also, one user-visible interaction may sometimes span multiple HTTP interactions, each of which falls below the threshold.

This monitoring facility is not expensive in terms of system resources and can be enabled in a production environment. Use a threshold value that is high enough to produce at most dozens rather than thousands of daily alerts.

Procedure

1. By default, this facility is enabled with a threshold of 1 second, or 1,000 milliseconds. Longer interactions that involve rules assembly are not reported.

2. To use a different threshold, update the alerts section of the prconfig.xml file, setting a longer or shorter desired threshold in milliseconds.

<env name="alerts/browser/interactionTimeThreshold/enabled"
      value="true" />
<env name="alerts/browser/interactionTimeThreshold/WarnMS"
      value="1000" />
<env name="alerts/browser/interactionTimeThreshold/excludeAssembly"      value="true" />

where warnMS is a threshold limit in milliseconds and excludeAssembly causes this facility to suppress logging when rule assembly has occurred during the current interaction.

3. Stop and restart the server.

4. Results are stored in the alert log. Each line produced includes the text:

HTTP interaction has exceeded the elapsed time alert threshold of Nnnn ms

As an alternative to the prconfig.xml file, you can use Dynamic System Settings to configure your application.
See How to create or update a prconfig setting.

The line also identifies the session ID, the response time, and portions of the input URL, and twelve elapsed time properties containing statistics from the Performance details display for the interaction. These statistics easier to review when you import the Alert log into Excel. PROJ-505 CODE-162 SEEMA 3/20/07 BUG-4065

• Elapsed time executing Declarative Rules — pxDeclRulesInvokedElapsed
• Elapsed time accessing non-rule-resolved instances from database — pxOtherIOElapsed
• Elapsed time retrieving rule-resolved Rules from the database — pxRuleIOElapsed
• Elapsed time retrieving Declarative Rules — pxDeclRulesLookupElapsed
• Elapsed time retrieving non-Rule database lists — pxOtherBrowseElapsed
• Elapsed time retrieving Rule database lists — pxRuleBrowseElapsed
• Elapsed time performing explicit database commit operations — pxCommitElapsed
• Elapsed time processing Storage Streams for Rule database lists — pxRuleBrowseFilterElapse
• Elapsed time performing Rule assembly — pxJavaAssembleElapsed
• Elapsed time compiling Rules — pxJavaCompileElapsed
• Elapsed time checking Java syntax — pxJavaSyntaxElapsed
• Elapsed time executing Connect Rules — pxConnectElapsed

For more information about these elapsed time statistics, see Performance tool — Full details display.

See also PDN article Understanding PEGA0001 Alerts.

	Alert log
	Understanding alerts

SysAdmin category