fix: rmq only join_cluster after reploy of 0 pod

We observed that running rabbitmqctl join_cluster during the restart of a RabbitMQ pod can lead to offline queues and potentially also missing consumers (though we have not been able to reproduce the issue, so take this information with a grain of salt). For this reason we will only run the join_cluster command for the mq-0 pod if and only if there is no data inside the /var/lib/rabbitmq/mnesia directory already. Other pods should be able to cluster with the mq-0 pod out of the box.