SMA-X/SMA-SM should have some sort of capability to react to situations, when pod(s) crashes.It should automatically run the "standard" scripts (delete pods, create a pod again etc.) without admin's own activity. (If a pod is already down, I assume no more harm can be done...) I'm sure R&D is able to say which of the troubleshooting scripts can be and are worth being automated. Of course, a log file describing the run scripts would be a nice addition.
Ideally, such an "automatic troubleshooting feature" would minimise a possible downtime and also admin's workload, if the standard tricks are already tried by system.