Support installing Agent Manager both standalone and on top of an existing OpenChoreo

[a5anka]

Context

Agent Manager (AM) and OpenChoreo (OC) currently ship as separate standalone installs. There are two gaps in how users can install AM today:

1. AM alongside OC. Users who want to run both side by side have to manually reconcile default ports, which is fragile — especially in local Docker setups where a restart can mean reinstalling one of them to recover. AM is also published as an OC ecosystem module (https://openchoreo.dev/ecosystem/item/?id=wso2-agent-manager), but installing the module onto an existing OC is a fully manual, step-by-step process. There's no quick install path, and no validation that the underlying OC has the capabilities AM needs (e.g. Thunder configuration).

2. AM on a VM. The current quick start installs AM locally and binds it to localhost, which doesn't work when you want to install AM on a VM with just Docker — the only practical option there is also a manual setup. There's no installation script aimed at the "VM with Docker" case.

Goal

AM should be installable across the common deployment scenarios with a comparable, scripted experience:

1. Standalone, local — today's quick start.
2. Standalone, on a VM with just Docker — a script that handles non-localhost binding and the rest of the Docker setup.
3. On top of an existing OC — a quick install path that adds AM to an existing OC deployment without manual configuration of underlying components.

Users should not have to drop down to manual, step-by-step setup for any of these.

Questions to align on

1. What should the "AM on a VM with Docker" install script cover — network/host config, TLS, persistence, anything else?
2. What should a "quick install AM on existing OC" flow look like end-to-end — what does it automate, and what does it expect the user to have already configured?
3. Should installation include capability validation, so AM fails fast if the underlying OC is missing something it needs?
4. How do we present the install modes to users so the right path is obvious for their environment?

Proposed next steps

• Add an installation script for AM on a VM with Docker
• Define a quick install path for AM on an existing OC
• Add capability validation between AM and OC at install time
• Document the supported install modes side by side