Add kernel crash dump functionality to all AMIs

[jsf9k]

🚀 Feature Proposal

I propose adding kernel crash dump functionality to all Linux AMIs that we build.

Motivation

In cisagov/guacamole-packer#26 we encountered a kernel bug that caused the Guacamole EC2 instances to lock up. The only way to get a clue as to what was happening was to enable kernel crash dumps, which I did manually. (Instructions for doing so are given in cisagov/guacamole-packer#26.) I think it makes a lot of sense to build this functionality on in all our AMIs, so if an instance locks up again we can get some information about what happened the first time it happens. Otherwise we would have to enable the functionality manually and wait for the lockup to happen again, like we did in cisagov/guacamole-packer#26.