fix(docker): restore ENTRYPOINT [start-container] — adding HEALTHCHECK had dropped it

Adding the HEALTHCHECK replaced the ENTRYPOINT line instead of sitting alongside
it, so the image fell back to the base php:8.4-fpm entrypoint (php-fpm alone):
supervisord/start-container never ran, nginx was down (port 80 unserved -> /api
404 via Traefik) and supervisorctl was absent (-> always UNHEALTHY). Restore the
entrypoint; HEALTHCHECK stays.

Dockerfile

@@ -187,3 +187,8 @@ EXPOSE 80
 HEALTHCHECK --interval=30s --timeout=10s --start-period=90s --retries=3 \
     CMD container-health
 
+# Boot supervisord (php-fpm + nginx + enabled workers) via start-container. This
+# MUST stay — without it the base php image's default entrypoint runs php-fpm
+# alone (no nginx -> port 80 unserved, no supervisord -> healthcheck fails).
+ENTRYPOINT ["start-container"]
+