Add strategy to restart downed member in the cluster

[ludovicc]

Select either
https://github.com/mbilski/akka-reasonable-downing or
https://github.com/guangwenz/akka-down-resolver

[ludovicc]

The strategy should ideally ensure that we can restart any downed process in the cluster where the following applications should be running: Woken, Woken validation, Portal backend.
If for exemple Woken is the leader of the cluster and must be restarted, then woken-validation and portal-backend should enter a failed state and request restart in their health checks. Marathon should then restart those processes, or they could even kill themselves and Marathon will have a easier job to restart them.