Through prconfig.xml
file settings, you can cause the system to add a line to the Alert log when a connector execution takes more than a set duration. This alert can help you identify that an external system is slow, or that data mapping of the request or response is slow.
When connector operations exceed a threshold time limit, the system adds a message to the alert log of this form, with additional details:
PEGA0020*The total connect interaction time has exceeded the specified threshold (ms)
with additional details.
For more information on debugging connectors, see Testing Services and Connectors, located on the Integration area of the PDN.
This facility is enabled by default. You can disable all or any of the four threshold values, or set larger or smaller threshold values.
1. Add these lines to the prconfig.xml
file:
<env name="alerts/connector/zzzzz/enabled" value="true" />
<env name="alerts/connector/zzzzz/threshold" value="nnnn" />
where zzzzz is one of these four keyword values:
totalRequestTime
— Enabled by default, with a default threshold of 1000 milliseconds. The actual duration corresponds to the Performance tool property pxConnectElapsed, which is labeled Elapsed time executing Connect Rules on the Performance full details display.outboundMappingTime
— Disabled by default. Measures the actual duration of the request data mapping. If you omit the second line, the default threshold is 500 milliseconds. The actual duration corresponds to the Performance tool property pxConnectOutMapReqTime, which is labeled Elapsed outbound mapping time on the Performance full details display.inboundMappingTime
— Disabled by default. Measures the actual duration of the response data mapping. If you omit the second line, the default threshold is 500 milliseconds. The actual duration corresponds to the Performance tool property pxConnectInMapReqTime. which is labeled Elapsed inbound mapping time on the Performance full details display.clientResponseTime
— Disabled by default. Measures the total elapsed time for the connector interaction, excluding both data mappings. If you omit the second line, the default threshold is 500 milliseconds. The actual duration corresponds to the Performance tool property pxConnectClientResponseElapsed, which is labeled Connect Client Response time on the Performance full details display:2. Stop and restart the server.
3. The system records connector executions that exceed the threshold value in the Alert log.
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.
For additional information on testing connectors, see Testing Services and Connectors, a document on the Integration pages of the PDN.
The PEGA0020 alert cannot be used to detect whether the external system is unavailable. If the external system is down or unreachable, no response is ever received. The totalRequestTime
, inboundMappingTime
, and clientResponseTime
alerts occur only if the external system sends a response.
This alert reports on executions of Connect EJB, Connect Java, Connect SOAP, Connect MQ, Connect JMS, and Connect MQ rules. They do not include executions of Connect SQL or Connect BPEL rules.
alert log, connector, master agents, prconfig.xml file | |
Understanding alerts
Performance tool — Full Details display |