I am considering moving some of my services from old school dedicated server to Kubernetes and containers.
Actually I have 6 dedicated servers running RabbitMQ each with 32 threads totalling 192 computing threads.
The dedicated servers are managing queues for 700 customers and my idea was to create 700 separate RabbitMQ containers to separate clients and improve scalability and reliability.
Reading about Kubernetes resources allocation I understand that: " The scheduler refuses to place a Pod on a node if the capacity check fails".
If I will create a Kubernetes cluster with the same computing power of 6 nodes with 192 threads, can I run 700 RabbitMQ container instances on it or will I need a lot more resources? Will the scheduler check require 700 threads ?
Then 1 server with 32 threads with 1 RabbitMQ instance seems to me to distribute load better than many smaller containers with limited cores resources, or can container share load among all cores on the Kubernetes host?