After rabbitmq starts, it occupies 1600m memory

[clcc2019]

After rabbitmq starts, it occupies 1600m memory？
kubernetes v1.20.15，rabbitmq image 3.8.34

[michaelklishin]

There are diagnostics tools that provide a breakdown of how the memory is used. Memory allocator settings can be changed, although without knowing what exactly uses most memory, this would be pointless.

FYI, RabbitMQ 3.8 is entirely out of support and 3.8 images are no longer produced by the community. Consider moving to 3.10 on Erlang 25.

[clcc2019]

There are diagnostics tools that provide a breakdown of how the memory is used. Memory allocator settings can be changed, although without knowing what exactly uses most memory, this would be pointless.

FYI, RabbitMQ 3.8 is entirely out of support and 3.8 images are no longer produced by the community. Consider moving to 3.10 on Erlang 25.

The default configuration I used and used 1.6G without the service running

[michaelklishin]

What I means was the rabbitmq-diagnostics command.

[michaelklishin]

3.8 is out of support => closing.

The system part is runtime allocators, so something in this environment causes it to allocate a larger than usual block ahead of time. We have seen this on CentOS Stream 9 hosts without any specifics to make a conclusion.