Fix unique_instance_id uniqueness by using system machine_id with fallback

What does this MR do and why?

Fixes the unique_instance_id uniqueness issue by implementing system-level identifier generation with fallback logic.

Problem

The current unique_instance_id implementation is not truly unique across GitLab instances. Analysis shows ~1.25K unique_instance_id values associated with multiple dim_installation_id records, indicating the same identifier is being generated by different instances.

Root Cause: Instances sharing the same db_key_base_truncated configuration (common in cloud deployments using DB snapshots or containerized environments) generate identical unique_instance_id values.

Solution

This MR implements a new system-level identifier generation approach:

1. Primary: Use /etc/machine-id when available (Linux with systemd)
2. Secondary: Use /proc/sys/kernel/random/boot_id when available.
3. Fallback: Use existing unique_instance_id logic for non-Linux systems

The new logic ensures better uniqueness by leveraging system-level identifiers that are unique per machine/container, rather than relying solely on database configuration.

References

Screenshots or screen recordings

Before	After

How to set up and validate locally

MR acceptance checklist

Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.

Related to #552160