We upgraded Apm agent from version 1.7.1 to 1.10. After release we discovered 100% CPU spikes on our production servers and slower performance. We removed that change and everything came back to normal. I cannot however provide any steps to reproduce, cause we cannot reproduce it locally.
Could new features in agent cause performance issues?
Do you have any test cases or load tests which check the performance?
Generally any idea how to find the cause or what could we do would be helpful.
APM Server version: 7.11.1 APM agent dotnet: 1.10 - Full net framework
I'd recommend upgrading to 1.11 first, to ensure that you're running with the latest version, and see if the same issue persists.
If the issue persists, can you share more about your environment and how the agent is configured and instantiated (if you're instantiating the agent) ? Be good to see if there is something obvious before spending time with profiling.
Profiling your application with a profiler such as
Should provide more detail around where time is spent. Ideally want to collect a dump on a production server, where the performance issue is exhibited.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.