RabbitMQ queue details parsing doesn't account for duplicated elements

[danielmarbach]

In some cases, RabbitMQ management API might return json elements with duplicated keys.

For example, with one of the customers it returned a duplicated memory element

{
    "filtered_count": 44,
    "item_count": 44,
    "items": [
        {
            "arguments": {
                "x-queue-type": "quorum"
            },
            "auto_delete": false,
            "consumer_capacity": 1,
            "consumer_utilisation": 1,
            "consumers": 1,
            "durable": true,
            "effective_policy_definition": {},
            "exclusive": false,
            "leader": "rabbit@7509bf5e7b2a",
            "members": [
                "rabbit@7509bf5e7b2a"
            ],
            "memory": 112564,
            "memory": 112564,

By accessing the json object dictionary, the dictionary is initialized and cannot contain multiple keys. JsonElement is more flexible in that regard.

Both this tool and ServiceControl lack proper testing infrastructure in place to reproduce these problems so I ended up manually loading a fake Json payload from a filestream into the parser.

[danielmarbach]

In Particular/Particular.EndpointThroughputCounter#848 I went down the rabbit hole of adding more coverage for the paging etc. which didn't exist but ran out of time to catch up here with things too.