Fix healthcheck host mismatch in reverse-proxy setups

[TylonHH]

Summary

• replace hardcoded curl -f http://localhost healthcheck with a host-aware request
• use Host: ${HTTP_SERVER_NAME:-localhost} so Laravel host trust checks match runtime domain config
• keep check local to container via http://127.0.0.1/

Problem

In reverse-proxy/swarm deployments, LinkStack may trust only the configured app domain host. A healthcheck request with Host: localhost can fail (e.g., 404/500), marking the container unhealthy and causing restart loops.

Change

HEALTHCHECK CMD curl -fsS -A "HealthCheck" -H "Host: ${HTTP_SERVER_NAME:-localhost}" http://127.0.0.1/ >/dev/null || exit 1

Impact

• keeps current behavior for default localhost setups
• fixes false-negative healthchecks when HTTP_SERVER_NAME is set
• avoids avoidable unhealthy restarts under orchestrators