Support Article
Kafka process runs for each node post application upgrade
SA-83628
Summary
Post upgrade to Pega 8.2.1, a Kafka process runs for each Pega JVM which consumes CPU storage. This results in performance degradation.
Error Messages
Not Applicable.
Steps to Reproduce
Unknown.
Root Cause
The Kafka process runs on Stream nodes and at least one Stream node is required for the Pega application to run. Unless otherwise specified, an 8.2.x node defaults to Background Processing, Search, WebUser, and Stream types.
Resolution
Perform the following local-change:To specify the node type, add the following JVM argument with a valid node type value such as WebUser:
DNodeType=WebUser
If the node is a Stream node, navigate to Dev Studio > Configure > Decisioning > Infrastructure > Services > Stream, and decommission the node. Restart it after applying the JVM argument.
Published August 19, 2019 - Updated December 2, 2021
Have a question? Get answers now.
Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.