I am trying to find average messages per month and average message size for my RabbitMQ cluster.

When I call curl -i -u user:pass http://FirstServer:port/api/vhosts I get back a structure that looks similar to this:

[
    {
        "cluster_state": {
            "rabbit@firstserver": "running",
            "rabbit@secondserver": "running",
            "rabbit@thirdserver": "running"
        },
        "description": "Default virtual host",
        "message_stats": {
            "ack": 19318651,
            "ack_details": {
                "rate": 49.2
            },
       ... parts removed for brevity...
            "deliver": 19349053,
            "deliver_details": {
                "rate": 25.4
            },
            "deliver_get": 19349053,
            "deliver_get_details": {
                "rate": 25.4
            },
       ... parts removed for brevity...
            "publish": 19318618,
            "publish_details": {
                "rate": 29.0
            }
       ... parts removed for brevity...
        },
        "messages": 224,
        "messages_details": {
            "rate": -30.2
        },
        "messages_ready": 17,
        "messages_ready_details": {
            "rate": 0.0
        },
        "messages_unacknowledged": 207,
        "messages_unacknowledged_details": {
            "rate": -30.2
        },
        "metadata": {
            "description": "Default virtual host",
            "tags": [
            ]
        },
        "name": "/",
        "recv_oct": 20573353083,
        "recv_oct_details": {
            "rate": 30245.4
        },
        "send_oct": 21329546408,
        "send_oct_details": {
            "rate": 27465.8
        },
        "tags": [
        ],
        "tracing": false
    }
]

Is there any documentation on this structure? Specifically, I have the following questions:

1. When it says rate, what are the units. (Per second, hour, month etc)
2. How often is the rate recalculated?
3. This url called the server FirstServer, but there are two others in the cluster. Are these stats for the whole cluster, or just for FirstServer?