Can we use a simple monolith mimir over multi nodes on a production environment ?

[sidkram]

I have been trying out mimir lately, and because of a few failures previously over deploying on kubernetes env as microservices, I had thought it should be better if I deploy mimir over ec2 nodes as mimir allows horizontal scalability. There are two ingester nodes and two querier nodes that I've been using, I've created a route53 private hosted zone one for ingester address and one for querier address that encapsulates two nodes ips each. This initially was working fine, but lately I've been noticing a few ingester crashes that is pulling the whole mimir cluster down. One node among the ingester group [.197] was used since the start and that had acted as the ingester node during the start. But the issue being, even being a 4 node cluster, when this .197 node goes down, the whole cluster fails. I need a few questions answered around this and a few suggestions if I'm missing out big here.

1. Can a monolith mimir ever be scaled to more than one node per service, is there even a allowance from the code to support that?
2. Does mimir or the config hold any logic for a specific node to be chosen for every service (in the case of monolith), the reason being, even if the ingester R53 address is mapped to two nodes, only the initial node [.197] gets the most traffic.