improvement of debug-on-exit.sh / fetch node OS logs

The script debug-on-exist.sh collects data from the kubernetes cluster when the deployment fails. However some logs missing like the OS status (/var/log/*, dump of some env etc.) and due the fact we build OS regarding our requirements (https://gitlab.com/sylva-projects/sylva-elements/diskimage-builder), having a snapshot of the OS could be sometime useful to understand why deployment fail (K8S logs are not enough sometimes). So it could be nice to add an invocation to the troubleshoot tool depending the target platform (Ubuntu=>sosreport, SuSE=>supportconfig) to collect data from all deployed nodes.