When deploying a large RabbitMQ cluster, the management plugin will impose a sensitive overhead on the cluster operations. In this recipe, we are showing how to relieve this overhead.
In order to experiment this recipe, you need a RabbitMQ cluster with at least two nodes. If they have the management plugin already installed, you need to remove it.
rabbitmq-plugins enable rabbitmq_management
rabbitmq-plugins enable rabbitmq_management_agent
After performing these steps, you will be able to monitor the entire cluster from the first node only. The other nodes will have the agents updating the status of the console of the first node, but you won't be able to access port 15672 on them.
Typically, you will install the full management plugin on a couple of nodes, such as frontend or management nodes, and the management agent plugin on the rest of them.