Type | Monitoring Metric | Unit | Description |
Basic Information | Connection Count | Count | Number of opened connections |
| Channel Count | Count | Number of opened channels |
| Queue Count | Count | Total number of available queues |
| Consumer Count | Count | Total number of online consumers |
| Message Backlog | Count | Total number of messages in the Ready state (accumulated but not delivered) |
| Inbound Public Network Bandwidth | Mbps | Inbound public network bandwidth |
| Outbound Public Network Bandwidth | Mbps | Outbound public network bandwidth |
| Recommended TPS Limit (Production + Consumption) | Count/s | Recommended TPS limit for the cluster when the mirror queue is not enabled |
| Total Opened Channels | Count | Total number of opened channels |
Production and Consumption | Production Confirmation Rate | Count/s | Rate that the broker confirms to return a packet after the client successfully generates a message |
| Messages Produced per Second | Count/s | Rate that the client generates messages |
| Unacknowledged Message Count | Count | Total number of messages that are delivered to but are not acknowledged by the consumer |
| Consumption Acknowledgment Rate | Count/s | Rate that the consumer acknowledges messages |
| Messages Consumed per Second | Count/s | Number of messages consumed per second, including scenarios in which autoAck = false and autoAck = true |
| Redelivery Rate | Count/s | Rate for redelivering messages to the consumer in a channel |
| Message Discard Rate | Count/s | Rate for dropping messages that are sent to an exchange with mandatory = false due to no matching route |
Type | Monitoring Metric | Unit | Description |
Basic Information | Connection Count | Count | Number of open connections |
| Channel Count | Count | Number of opened channels |
| Queue Count | Count | Total number of available queues |
| Consumer Count | Count | Total number of online consumers |
| Message Backlog | Count | Total number of messages in the Ready state (accumulated but not delivered) |
| CPU Usage | % | CPU usage of a node |
| Memory Usage | % | Memory usage of a node |
| Disk Usage | % | Disk usage of a node |
Production and Consumption | Production Confirmation Rate | Count/s | Rate that the broker confirms to return a packet after the client successfully generates a message |
| Messages Produced per Second | Count/s | Rate that the client generates messages |
| Unacknowledged Message Count | Count | Total number of messages that are delivered to but are not acknowledged by the consumer |
| Consumption Acknowledgment Rate | Count/s | Rate that the consumer acknowledges messages |
| Messages Consumed per Second | Count/s | Number of messages consumed per second, including scenarios in which autoAck = false and autoAck = true |
| Redelivery Rate | Count/s | Rate for redelivering messages to the consumer in a channel |
| Message Discard Rate | Count/s | Rate for dropping messages that are sent to an exchange with mandatory = false due to no matching route |
Type | Monitoring Metric | Unit | Description |
Basic Information | Consumer Count | Count | Total number of online consumers |
| Message Backlog | Count | Total number of messages in the Ready state (accumulated but not delivered)
++ |
Production and Consumption | Production Confirmation Rate | Count/s | Rate that the broker confirms to return a packet after the client successfully generates a message |
| Messages Produced per Second | Count/s | Rate that the client generates messages |
| Unacknowledged Message Count | Count | Total number of messages that are delivered to but are not acknowledged by the consumer |
| Consumption Acknowledgment Rate | Count/s | Rate that the consumer acknowledges messages |
| Messages Consumed per Second | Count/s | Number of messages consumed per second, including scenarios in which autoAck = false and autoAck = true |
| Redelivery Rate | Count/s | Rate for redelivering messages to the consumer in a channel |
| Message Discard Rate | Count/s | Rate for dropping messages that are sent to an exchange with mandatory = false due to no matching route |
Type | Monitoring Metric | Unit | Description |
Basic Information | Consumer Count | Count | Total number of online consumers |
| Message Backlog | Count | Total number of messages in the Ready state (accumulated but not delivered) |
Production and Consumption | Unacknowledged Message Count | Count | Total number of messages that are delivered to but are not acknowledged by the consumer |
| Consumption Acknowledgment Rate | Count/s | Rate that the consumer acknowledges messages |
| Redelivery Rate | Count/s | Rate for redelivering messages to the consumer in a channel |
Icon | Note |
| Click it to view the monitoring metrics on a YoY basis. YoY, MoM, and custom date are supported. |
| Click it to refresh and obtain the latest monitoring data. Monitoring data can be refreshed at intervals of 30 seconds, 5 minutes, 30 minutes, and 1 hour. |
| |
| After it is selected, legend information can be displayed on the chart. |
Metric | Dimension | Suggested Alarm Configuration | Description |
Disk Usage (%) | Node | Statistical period of 1 minute, > 80%, and for 5 consecutive data points. Alarm once every 30 minutes | High disk usage may result in insufficient disk space on the node to accommodate messages assigned to it. As a result, messages cannot be written to the disk. You are advised to promptly clear data or scale out the cluster when the average disk usage exceeds 80%. |
Memory Usage (%) | Node | Statistical period of 1 minute, > 50%, and for 5 consecutive data points. Alarm once every 30 minutes | High memory usage blocks message production. You are advised to accelerate consumption, apply flow control to production, or scale out the cluster when the memory usage exceeds 50%. |
CPU Usage (%) | Node | Statistical period of 1 minute, > 70%, and for 5 consecutive data points. Alarm once every 30 minutes | High CPU usage affects the message production speed. You are advised to scale out the cluster when the CPU usage exceeds 70%. |
Message backlog (Count) | Node | Statistical period of 5 minutes, > expected message backlog for the business, and for 5 consecutive data points. Alarm once every 30 minutes | An excessive accumulation of messages causes a rapid disk usage increase of the broker node. As a result, other messages cannot be received, and scale-out is needed. |
Was this page helpful?