on :all, in: :sequence, wait: 15 do
# This takes all servers, in sequence and waits 15 seconds between
# each server, this might be perfect if you are afraid about
# overloading a shared resource, or want to defer the asset compilation
# over your cluster owing to worries about load
end