How to capture a coredump in OCP of a process which is running inside a container ?

Issue

• Any c++ (or otherwise "native") application needs to have core dump available for proper troubleshooting of crashes. In OCP the default 'core_pattern' points at 'systemd-coredump'. This effectively disables the feature.
• We have outlined few unsatisfying solutions:

      1. Have an 'init-container' which changes the 'core_pattern'. This requires privileged security context 
         and respective SCC.
      2. Change the 'core_pattern' cluster wide by using 'MachineConfig' - this is often not acceptable in multi 
         tenant environment whenever the cluster is managed by another party.

• Does OCP have a standard way to allow for proper core file collection? 'Systemd-coredump' is not viable as 'systemd' usually is not part of the containers for good reasons.
• The impact is high because it either cripples our troubleshooting capabilities or causes friction with our clients who usually own the clusters.