We have 4 magento servers services an application. When I installed the oneagent and restarted the nginx and php processes, the PHP processes go from using ~100m of memory each to using multiple GB of memory, and the application stops responding. Turning oneagent off and restarting application stops the problems and the app works again.
I do have a support ticket on this being worked, but was just curious if anyone else has seen anything remotely similar on magento/PHP-FPM regarding runaway consumption of resources.
Solved! Go to Solution.
Just so anyone else that might have this issue in the future, it was fixed by turning off the class search functionality for the php agent. Development is still researching why that causes so much overhead in this particular setup.
Specifically we set debugDisableClassSearchPhp debug flag to true which disabled the class search functionality within our PHP special agent