Allow HA scale to be configurable beyond 2 replicas

### Description

A customer plans to use the HA option to provide scalability of their NR instance. They foresee a need, in the next few months, to scale beyond the current limit of 2 replicas.

This issue is to track how we address this requirement.

The quick option is to allow the user to set the replica count for themselves (with some reasonable limit).

The harder option is to look at auto-scaling; how much of that can be delegated to k8s, or would we need to manage it.

We also need to consider billing model implications. Currently HA spins up a second replica - so uses double the resources - without changing the cost of the instance. If we allow an instance to spin up 10 replicas... the resource cost will be much greater. It feels like that should have a cost implication - need to figure out options around that (on both business and technical side).

Customer
- https://app-eu1.hubspot.com/contacts/26586079/record/0-1/27210478021

### Which customers would this be available to

Enterprise Tier Only (EE)

### Have you provided an initial effort estimate for this issue?

I can not provide an initial effort estimate

Start by tracing how HA currently creates its second replica and where Kubernetes could receive replica or autoscaling settings; the issue names no files or tests, so repository research is needed. Compare fixed user-configurable limits with delegated autoscaling and document the enterprise-only billing implications; done requires an agreed technical and pricing approach.

Written by the indexing model from the issue text.