Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Accessing Prometheus metrics

Amazon MQ now supports Prometheus metrics for Amazon MQ for RabbitMQ brokers. Prometheus metrics enable you to integrate broker observability into your existing monitoring infrastructure, giving you a unified view of broker performance alongside your other services. With Prometheus metrics, you can set up fine-grained alerting and dashboards to proactively detect and respond to issues in your messaging workloads.

Starting with RabbitMQ 4.2, Amazon MQ for RabbitMQ supports Prometheus metrics, allowing you to scrape broker metrics using the Prometheus monitoring system. The following endpoints are supported:

The /metrics/per-object endpoint is not supported.

For more information about the metrics exposed by each endpoint, see Prometheus metric in the RabbitMQ documentation.

Prometheus metrics vs CloudWatch metrics

Amazon MQ for RabbitMQ exposes metrics through both Prometheus endpoints and CloudWatch. While both provide visibility into broker health, they differ in scope and usage.

The Prometheus endpoints expose a richer set of aggregated metrics about RabbitMQ broker health, covering a broader range of broker internals such as connection churn, channel activity, queue and exchange statistics, and Raft consensus metrics. These are suited for integration with existing Prometheus-based monitoring infrastructure and fine-grained alerting.

CloudWatch metrics are a curated subset of broker metrics obtained from the Prometheus endpoints. For a full list of available CloudWatch metrics, see Available CloudWatch metrics for Amazon MQ for RabbitMQ brokers.

In CloudWatch, metrics are always aggregated with an interval of at least 60 seconds before visualization. In contrast, Prometheus exposes raw metric datapoints, and dashboard solutions like Grafana visualize individual datapoints without aggregation by default. As a result, visualizations of the same metric can diverge between CloudWatch and Prometheus depending on the statistic used in CloudWatch

Obtaining and accessing the Prometheus endpoints

You can obtain the Prometheus endpoint for your Amazon MQ for RabbitMQ broker using the Amazon Web Services Management Console or the Amazon CLI.

Amazon MQ for RabbitMQ Prometheus support uses the same authentication scheme as the broker. For more information about the supported authentication methods, see Amazon MQ for RabbitMQ Authentication and Authorization. To learn how to configure authentication in Prometheus, see http_config in the Prometheus documentation.

Prometheus configuration best practices

Sample scraping configuration

The following sections provide sample Prometheus scraping configurations for Amazon MQ for RabbitMQ. Replace <broker-prometheus-endpoint> with your broker's Prometheus endpoint hostname, and <username> and <password> with your broker credentials.

Recommended configuration

The following configuration is recommended for most use cases. Scraping the /metrics endpoint provides well-aggregated metrics about overall cluster health, giving you a clear view of broker performance without the overhead of detailed metric collection.

global:
  scrape_interval: 60s

scrape_configs:
  - job_name: 'rabbitmq-aws-cluster'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics'
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'

Detailed metrics configuration

The following configuration scrapes additional detailed metric families for deeper observability into specific broker components.

global:
  scrape_interval: 60s

scrape_configs:
  - job_name: 'rabbitmq-connection-churn'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['connection_churn_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-ra'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['ra_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-queue'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['queue_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-exchange'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['exchange_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-connection'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['connection_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-channel'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['channel_metrics']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'
  - job_name: 'rabbitmq-exchange-count'
    scheme: https
    basic_auth:
      username: <username>
      password: <password>
    metrics_path: '/metrics/detailed'
    params:
      family: ['exchange_names']
    static_configs:
      - targets:
        - '<broker-prometheus-endpoint>:16001'
        - '<broker-prometheus-endpoint>:16002'
        - '<broker-prometheus-endpoint>:16003'