[Questions] Is management metrics collection deprecated for the UI only or is the API deprecated as well?

[rsperl]

Community Support Policy

• I have read RabbitMQ's Community Support Policy
• I run RabbitMQ 4.x, the only series currently covered by community support
• I promise to provide all relevant information (versions, logs from all nodes, rabbitmq-diagnostics output, detailed reproduction steps)

RabbitMQ version used

4.0.9

Erlang version used

26.2.x

Operating system (distribution) used

rhel9

How is RabbitMQ deployed?

Other

rabbitmq-diagnostics status output

See https://www.rabbitmq.com/docs/cli to learn how to use rabbitmq-diagnostics

n/a

Logs from node 1 (with sensitive values edited out)

See https://www.rabbitmq.com/docs/logging to learn how to collect logs

n/a```
</details>


### Logs from node 2 (if applicable, with sensitive values edited out)

See https://www.rabbitmq.com/docs/logging to learn how to collect logs
<details>

n/a

</details>


### Logs from node 3 (if applicable, with sensitive values edited out)

See https://www.rabbitmq.com/docs/logging to learn how to collect logs
<details>

n/a

</details>


### rabbitmq.conf

See https://www.rabbitmq.com/docs/configure#config-location to learn how to find rabbitmq.conf file location
<details>

n/a

</details>


### Steps to deploy RabbitMQ cluster

n/a

### Steps to reproduce the behavior in question

n/a

### advanced.config

See https://www.rabbitmq.com/docs/configure#config-location to learn how to find advanced.config file location
<details>

n/a

</details>


### Application code

<details>

```python
n/a

Kubernetes deployment file

n/a

What problem are you trying to solve?

I see that management_metrics_collection is deprecated, and when I turn it off, I can see that the metrics are gone from the UI as expected. Are there specific API endpoints that are also deprecated as part of this or is it only the UI that is affected?

[rsperl]

Ok - answered my own question. Some existing endpoints change, such as /api/queues no longer returns queue related metrics.

[michaelklishin]

@rsperl RabbitMQ 4.0.9 is not covered by community support any more.

GET /api/queues does return queues metrics if the management plugin isn't configured to not return them:

[
  {
    "arguments": {
      "x-queue-type": "classic"
    },
    "auto_delete": false,
    "consumer_capacity": 0,
    "consumer_utilisation": 0,
    "consumers": 0,
    "durable": true,
    "effective_policy_definition": {},
    "exclusive": false,
    "internal": false,
    "internal_owner": false,
    "memory": 14048,
    "message_bytes": 0,
    "message_bytes_paged_out": 0,
    "message_bytes_persistent": 0,
    "message_bytes_ram": 0,
    "message_bytes_ready": 0,
    "message_bytes_unacknowledged": 0,
    "messages": 15,
    "messages_details": {
      "rate": 0.0
    },
    "messages_paged_out": 0,
    "messages_persistent": 15,
    "messages_ram": 1,
    "messages_ready": 15,
    "messages_ready_details": {
      "rate": 0.0
    },
    "messages_ready_ram": 1,
    "messages_unacknowledged": 0,
    "messages_unacknowledged_details": {
      "rate": 0.0
    },
    "messages_unacknowledged_ram": 0,
    "name": "qq.1",
    "node": "rabbit@W4P56M04DC",
    "reductions": 7388,
    "reductions_details": {
      "rate": 0.0
    },
    "state": "running",
    "storage_version": 2,
    "type": "classic",
    "vhost": "/"
  }
]

Some metrics were dropped in 4.0 because in total there were at least 60 of them per queue, and GET /api/queues without any pagination parameters was a very common problem to see (and the endpoint was abused to fetch one metric of one queue quite often on top).

The recommendation is to use Prometheus and Grafana.

The metrics part of the management plugin is fairly likely to get undeprecated because no matter what we do, some users will never adopt a dedicated monitoring system, so management UI metrics is the only thing they'd ever have. But no decision has been made on that.

[michaelklishin]

More metrics will be dropped for classic queues over time because some of them made sense with previous generations of CQ storage but not so much any more. The idea of transient messages is very likely to go away in the future because they cause a lot more problems than they solve => related metrics will be dropped as a result.