Cross-posted from Erlang forums…
We have an AWS based Kuberentes orchestrated container deployment. The Erlang VM is the only application within that container, and the container is provisioned with 16 cores, of which I’m sure half are hyper-threaded.
Now there is no real evidence of performance issues, but the VM is configured with “factory defaults”. My question is what, if any, changes should be applied to the VM configuration to run optimally in such an environment? I’m specifically thinking of things like scheduler bind type and scheduler busy wait?