REST-triggered DB interactions exceed desired timeouts
SummaryUsers' Connect-REST responses are extremely slow and system hangs when the JSON responses are large.
Even after applying the HFix-23709 and priming the JSON Property cache by making Connect-REST calls, the same performance issue as described above continue to exist.
A sample ALERT entry with PAL data is pasted below.
User have also captured DBTrace for the above Connect-REST call and sees 38,677 occurrences of the SQL query being fired below:
SELECT pyPropertyName AS "pyPropertyName", pyPageClass AS "pyPageClass", pyPropertyMode AS "pyPropertyMode" FROM PEGA_RULES.pr4_rule_property "property" WHERE (pxObjClass = 'Rule-Obj-Property') AND (pyClassName = 'Class-Name') AND (pyRuleAvailable = 'Yes')
Error Messages2016-10-11 19:02:27,207 GMT*8*PEGA0020*82320*1000**pxTotalReqTime=82.32;pxAlertCount=1;pxRulesUsed=12;pxRulesExecuted=4;pxTrackedPropertyChangesCount=38,677;pxRDBIOCount=38,677;pxTotalReqCPU=59.88;pxRDBWithoutStreamCount=38,677;pxRDBIOElapsed=39.27;*RULE-CONNECT-REST CLASSNAME RULENAME #20161011T185820.104 GMT has exceeded the threshold of 1,000 ms: 82,320 ms;pxConnectOutMapReqElapsed=1;pxConnectClientResponseElapsed=8861;pxConnectInMapReqElapsed=33630;*
Steps to ReproduceInvoke the Connect-REST calls.
Root CauseA defect in Pegasystems’ code or rules.
High amount of time spent in performing JSON Property lookup (pxRDBIOElapsed) and mapping the incoming JSON response to the clipboard properties (pxConnectInMapReqElapsed).
Published October 20, 2016 - Updated November 8, 2016