Please help to find the root cause for "child 97041 exited on signal 11 (SIGSEGV)"

[Offerel]

Today I installed version 10.0.2 of Recognize. I had a lot of “502 Gateway Error” messages during this. For some reason, PHP-FPM stopped responding. I assume, this is not really the fault of Recognize, but it manifest during installation. I assume the server probably ran out of resources. In the PHP-fpm log file, I found a lot of error messages similar to “WARNING: [pool www] child 97041 exited on signal 11 (SIGSEGV) after 175.908460 seconds from start."

Restarting php8.4-fpm solved the problem for the moment. I would like to avoid such situations in the future and only restart PHP-FPM automatically in an emergency. However, it would be important for me to fix the root cause of the problem first.

Does anyone here know of a way to automatically create a kind of report/log in this case that shows me the top 20 memory hogs and possibly the same for the CPU load?

Presumably, PHP with its child processes is the memory hog, but I would like to see that directly in a log file. I know how to write the processes to a file with the data and format I want. But how do I trigger the creation of the file when the error “WARNING: [pool www] child 97041 exited on signal 11 (SIGSEGV)” appears in the log file?

BTW: The server is Debian 13 with nginx 1.26.3 and php8.4-fpm, mysql 11.8.2