Getting more metrics returned than I expected

[shaunokeefe]

G'day. I'm having some trouble with performance so I'm looking at limiting the exporter a smaller subset of the default metrics. Currently I'm looking at only exporting rabbitmq_queue_messages_ready and rabbitmq_exchange_messages_published_out_total. I've got the following configuration:

 {rabbit, [
        ...
    ]},
    {prometheus, [
        {collectors, [
           prometheus_rabbitmq_queues_collector,
           prometheus_rabbitmq_exchanges_collector
        ]},
        {rabbitmq_exporter, [
            {path, "/metrics"},
            {queue_messages_stat, [
                rabbitmq_queue_messages_ready
            ]},
            {exchange_messages_stat, [
                rabbitmq_exchange_messages_published_out_total
            ]}
        ]}
    ]},
    {rabbitmq_management, [
      {listener, [
        {port, 15672},
        {ssl, false}
      ]}
    ]},

This excludes metrics from collectors that aren't prometheus_rabbitmq_queues_collector and prometheus_rabbitmq_exchanges_collector collectors (i.e. no mnesia table metrics etc), but I was hoping that the settings I had for exchange_messages_stat and queue_messages_stat would further limit the output from polling /metrics to the two metrics I was after. Instead it looks like pretty much all available metrics for the two collectors are returned when I query the metrics endpoint.

Any idea what I'm doing wrong?

[deadtrickster]

Hi, I think naming might be the problem here: please try something from here:

prometheus_rabbitmq_exporter/src/prometheus_rabbitmq_exporter_config.erl

Line 13 in 1c59805

-define(DEFAULT_QUEUE_MESSAGES_STAT, [messages_published_total,