It is possible to configure a memory high watermark on RabbitMQ to define memory thresholds using the memoryHighWatermark.* parameters. This can be done in any of the following way:
Set an absolute limit of RAM to be used on each RabbitMQ node, as shown in the configuration example below:
memoryHighWatermark.enabled="true"
memoryHighWatermark.type="absolute"
memoryHighWatermark.value="512MB"
Set a relative limit of RAM to be used on each RabbitMQ node. To enable this feature, define the memory limits at pod level too. An example configuration is shown below:
memoryHighWatermark.enabled="true"
memoryHighWatermark.type="relative"
memoryHighWatermark.value="0.4"
resources.limits.memory="2Gi"