[RFE] include an action for joining a cluster

There is a lot of effort required to deal with the use case of an unhealthy RabbitMQ node that needs to be refreshed (database regenerated) and rejoined to the cluster.

Once the unit is paused (action 'pause') and the node removed from the cluster (action 'forget-cluster-node'), on the node's machine:

1. Move the current database out of the way:

$ sudo mv /var/lib/rabbitmq/{mnesia,mnesia.bak}

2. Start the server in standalone (unclustered) mode

$ sudo rabbitmq-server -detached

3. Stop the existing erlang VM

$ sudo rabbitmqctl stop_app

4. Rejoin the unhealthy node to the cluster by referencing an existing cluster member:

$ sudo rabbitmqctl join_cluster rabbit@juju-64dabb-2-lxd-0

5. Start a refreshed erlang VM

$ sudo rabbitmqctl start_app

6. Stop the RabbitMQ service in the current manually-started environment in order for it to be managed by Juju:

$ sudo rabbitmqctl stop

The unit can then be resumed (action 'resume').

FWIW, the mysql-innodb-cluster charm has an action called 'rejoin-instance'.