Causes of Excessive Transaction Traces for Java Agents and how to contain them.

Products

• CA Application Performance Management Agent (APM / Wily / Introscope)
• INTROSCOPE

Issue/Introduction

  This knowledge document was suggested by APM Sustaining Engineering and supplements https://ca-broadcom.wolkenservicedesk.com/external/article?articleId=10688 -- Ways to Contain/Control Transaction Traces.

  What are some of the different ways that too many transaction traces can be produced for a Java Agent? If this happens, what short-term measures should be taken?

Environment

All Supported APM releases

Resolution

  The common scenarios for too many transaction traces are below as well as suggested short-term measures to control the situation.

1. Excessive TTs in backend Agent triggered by upstream sampling. – disable/reduce upstream sampling.
2. Excessive TTs triggered by Auto Tracing in 10.x Agents – disable Auto tracing.
3. Excessive TTs in backend Agent triggered by upstreaming Auto Tracing – disable Auto tracing CPTT.
4. Excessive TTs triggered by ERRORs – disable/reduce Errors produced.
5. Excessive TTs in backend Agent triggered by upstream Agent due to tail filter. – disable tail filter propagation.
6. Excessive TTs in backend Agent triggered by upstream Agents for unknown reasons – disable HTTP/SOAP/JMS/MQ header insert/read operations.
7. Excessive TTs triggered by Field Packs – remove/upgrade Field Pack
8. Excessive TTs for unknown reasons – disable global TT.