Is there any document & any experience with 1 node with 4 vcpu & 64GB memory, that install with 2 instance or 3 instance for production use? Or is there best practice for 1 node with 4 vcpu & 64GB with how much instant can be install?
There’s no general recommendation as it might depends on the load and traffic each single instance needs to handle. Best practices would be to do proper testing with the estimated load in test environment and then then decide for production.