Monitoring & Management

Web UI

Overview

Ready: 已經發佈到 Queue 的訊息量
Unacked: 消費端已接收但尚未 Ack 的訊息量
Total: Ready + Unacked

CLI

rabbitmq-diagnostics

Online Resource Utilization

rabbitmq-diagnostics observer

RabbitMQ Version

[root@tpeeaprmq98 ~]# rabbitmq-diagnostics server_version
Asking node rabbit@tpeeaprmq98 for its RabbitMQ version...
3.10.7

Check the listener ports

rabbitmq-diagnostics -s listeners

rabbitmqctl

List the queues

rabbitmqctl -p <vhost-name> list_queues name state durable arguments policy

rabbitmqctl -qs -p <vhost-name> list_queues name > queue_names.lst

User Management

其他 tag 資訊: Management Plugin — RabbitMQ

# List all users
rabbitmqctl list_users

# Create a new user
rabbitmqctl add_user "eapuser"
rabbitmqctl add_vhost "eap_server"
rabbitmqctl set_permissions -p "eap_server" "eapuser" ".*" ".*" ".*"
rabbitmqctl set_user_tags eapuser monitoring  # for web login only

# Remove tag 'monitoring' from user if needed
rabbitmqctl set_user_tags eapuser ""

# Tag the user with "administrator" for full management UI and HTTP API access
rabbitmqctl set_user_tags username administrator

## Verifying the permission
# => Listing permissions for vhost "/" ...
# => user    configure   write   read
# => user2   .*  .*  .*
# => guest   .*  .*  .*
# => temp-user   .*  .*  .*
rabbitmqctl list_permissions --vhost /
rabbitmqctl list_permissions --vhost gw1

# Revoke user access
rabbitmqctl delete_user 'username'

Connections

rabbitmqctl list_connections
rabbitmqctl list_connections user,peer_host,peer_port,channels,state

Recreate the virtual host

rabbitmqctl delete_vhost <my-vhost-name>
rabbltmqctl add_vhost <my-vhost-name>

Reset the RabbitMQ Node

The broker drops all virtual hosts, queues, exchanges, and non-administrative users.

rabbitmqctl stop_app
rabbitmqctl reset
rabbitmqctl start_app

Force Reset the RabbitMQ Node

如果 Cluster 損壞且無法恢復運作時，可以嘗試強制重設 Node。

rabbitmqctl stop_app
rabbitmqctl force_reset
rabbitmqctl start_app

Suspend all listeners and prevent new client connections

暫停 listener 後，可以使用指令 ss -ltpn 檢查所有的 listener port 是否關閉。

既有的連線不會受到影響，不過新的連線將無法建立。

# For current node
rabbitmqctl suspend_listeners

# suspends listeners on node rabbit@node2.cluster.rabbitmq.svc: it won't accept any new client connections
rabbitmqctl suspend_listeners -n rabbit@node2.cluster.rabbitmq.svc

# For current node, to resume all listeners on a node and make it accept new client connections again
rabbitmqctl resume_listeners
# resumes listeners on node rabbit@node2.cluster.rabbitmq.svc: it will accept new client connections again
rabbitmqctl resume_listeners -n rabbit@node2.cluster.rabbitmq.svc

rabbitmqadmin

Basic Operation

# List queus
rabbitmqadmin list queues
rabbitmqadmin -H <RabbitMQ-Server-IP> -u <username> -p <password> -V <vhost-name> list queues

# Add a queue with optional parameters
rabbitmqadmin declare queue name=<my-new-queu> durable=true auto_delete=true

Remove multiple queues

rabbitmqadmin -f tsv -q list queues name > q.txt
while read -r name; do rabbitmqadmin -q delete queue name="${name}"; done < q.txt

Connections

rabbitmqadmin -H <RabbitMQ-Server-IP> -u <username> -p <password> -V <vhost-name> list connections name

# Close multiple connections without any channels
rabbitmqadmin -f tsv -q connections name channels | awk -F "\t" '($2 < 1) {print $1}' | tee conn_noChannels.lst
while read -r conn;do rabbitmqadmin close connection name="${conn}"; done < conn_noChannels.lst

Monitoring

Health-Check

rabbitmq-diagnostics check_running
rabbitmq-diagnostics ping

Queue State

rabbitmqctl list_queues name state

Cluster Status

rabbitmqctl cluster_status

Monitor with Prometheus

RabbitMQ Configuration

# Enable the plugin rabbitmq_prometheus
rabbitmq-plugins enable rabbitmq_prometheus

# To confirm that RabbitMQ now exposes metrics in Prometheus format
curl -s localhost:15692/metrics | head -n 10

Prometheus Configuration

/etc/hosts:

<node1.ip.addr>	 rmq01
<node2.ip.addr>	 rmq02
<node3.ip.addr>	 rmq03

prometheus.yml:

scrape_configs:
  - job_name: rabbitmq

    # Override the global default and scrape targets from this job every 5 seconds.
    scrape_interval: 15s
    
    static_configs:
      - targets: ['rmq01:15692', 'rmq02:15692', 'rmq03:15692']