Hi, I tried your method to trace system calls invoked by a container(more specifically, Nginx), but failed since that Nginx has multiple processes(1 master and N worker). The master process always has pid 1, but it’s hard to predict the pid of worker processes. Thus, `CMD [“strace”, “-p”, “1”]` does not work well in such cases.
Do you have any suggestion for me to debug on multiple-processes-application?
Btw, before reading your post, I always kept debugging container in host(not a separate container), I’d like to say it’s much more straight, except that you have to find out which is the container process.