Support Article
Poor Performance/Hung PRPC Node Instance
SA-12050
Summary
PEGA 7.1.7 users complain of generally poor system performance. Intermittantly, the PRPC node will become non-responsive, requiring a shutdown/restart.
The Pega Sys Admin has reviewed the Pega logfile, and concluded that thread references to a third party application called HazelCast are to blame for the hung threads/poor performance.
Error Messages
No error messages are displayed on screen. However, the PRPC log does contain a thread dump detailing hung threads. The thread dump is too voluminous to include in this explanation. However, the analysis of this thread dump is provided following.
Steps to Reproduce
No specific steps were found to reproduce this error condition. Poor performance would seemingly afflict the users throughout the business day - from initial login all the way to user logout.
Root Cause
Examination of the provided logfiles reveals a thread dump. The thread dump root cause was a hung thread, waiting on a CONNECT-SOAP operation to complete. Examination of the CONNECT-SOAP operation revealed the Pega ALERT EXCP0001 was being thrown, with an error condition indicating that the CONNECT-SOAP had failed due to a problem with the SOAP Service.
Resolution
Advised systems administrator to contact their SOAP service provider, and discuss Service Level Agreements regarding the service. The customer has since confirmed that the service is failing, and the administrator has confirmed root cause. Service is being repaired/tuned to accomodate the increased load of calls.
Published July 14, 2015 - Updated October 8, 2020
Have a question? Get answers now.
Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.