RFE: support mult. rgw frontends per node

### Problem

On high core-count nodes, a single RGW daemon doesn;t always scale throughput linearly. In throughput testing we see per-daemon contention / limited core utilisation. Running multiple RGW daemons per physical node is a common way to scale RGW capacity, but MicroCeph currently doesnt support his

### Proposal

Add support to run N RGW daemons per node, each with its own frontend bind address/port.

* Each instance binds to a unique port (e.g. base port + index).
* Provide a predictable naming scheme for daemons (e.g. `rgw..`).
* TLS material should be shared across RGW daemons

### UI / frontend ideas

Could req. the user run `microceph enable rgw` mult. times with distinct ports, or provide a `microceph enable rgw --num N` flag where N daemons would be spawned with ports starting from a base port

No files, tests, or entry points are named. Start by locating the RGW enable path and how daemon ports, names, and TLS material are represented; done means multiple RGW daemons can run per node with unique predictable ports and shared TLS material.

Written by the indexing model from the issue text.