Amazon MQ for RabbitMQ: Disk limit alarm - Amazon MQ
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).

Amazon MQ for RabbitMQ: Disk limit alarm

Disk limit alarm is an indication that the volume of disk used by a RabbitMQ node has decreased due to a high number of messages not consumed while new messages were added. RabbitMQ will raise a disk limit alarm when the broker's free disk space, identified by Amazon CloudWatch metric RabbitMQDiskFree, reaches the disk limit, identified by RabbitMQDiskFreeLimit. RabbitMQDiskFreeLimit is set by Amazon MQ and has been defined considering the disk space available for each broker instance type.

An Amazon MQ for RabbitMQ broker that has raised a disk limit alarm will become unavailable for new messages being published. When running RabbitMQ in a cluster, the disk alarm is cluster-wide. If one node goes under the limit, all other nodes will block incoming messages. Due to the lack of disk space, your broker might also experience other issues that complicate diagnosis and resolution of the alarm.

Amazon MQ will not restart a broker experiencing a disk alarm and will return an exception for RebootBroker API operations as long as the broker continues to raise the alarm.

Note

You cannot downgrade a broker from an mq.m5 instance type to an mq.t3.micro instance type. If you wish to downgrade, you must delete your broker and create a new one.

Diagnosing and addressing disk limit alarm

Amazon MQ enables metrics for your broker by default. You can view your broker metrics by accessing the Amazon CloudWatch console, or by using the CloudWatch API.MessageCount is a useful metric when diagnosing the RabbitMQ disk limit alarm. Messages are stored in memory until they are consumed or discarded. A high message count indicates overutilization of disk storage and can lead to a disk alarm.

To diagnose the disk limit alarm, use the Amazon MQ Management Console to:

  • Consume messages published to the queues.

  • Purge messages from queues.

  • Delete the queues from your broker.

Note

It may take up to several hours for the RABBITMQ_DISK_ALARM status to clear after you take the required actions.

To prevent the disk limit alarm from reoccurring, you can upgrade your host instance type to an instance with additional resources. For information on how to update your broker's instance type see UpdateBrokerInput in the Amazon MQ REST API Reference.