fix(svc-prx-openresty): swarm-aware handler — exec reload + container_address

Replace bare-name docker restart with container_address-resolved
openresty -s reload so the handler works on swarm tasks (named
<stack>_<svc>.<n>.<id>) without dropping listeners on 80/443. Apply
the same container_address fix to the MODE_DEBUG logs collection.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

roles/svc-prx-openresty/handlers/restart_and_validate.yml

@@ -10,9 +10,10 @@
     ('no such container' not in ((openresty_test_pre.stderr | default('')) | lower)) and
     ('is restarting' not in ((openresty_test_pre.stderr | default('')) | lower))
 
-- name: Restart OpenResty container
-  ansible.builtin.command: >
-    container restart {{ OPENRESTY_CONTAINER }}
+- name: Reload OpenResty configuration (in-place, preserves listeners)
+  ansible.builtin.shell: >
+    container exec {{ lookup('container_address', 'svc-prx-openresty', 'openresty') }}
+    openresty -s reload
   register: openresty_restart
   changed_when: true
   failed_when: >
@@ -32,8 +33,8 @@
       failed_when: false
       ignore_errors: true
 
-    - ansible.builtin.command: >
-        container logs --tail=200 {{ OPENRESTY_CONTAINER }}
+    - ansible.builtin.shell: >
+        container logs --tail=200 {{ lookup('container_address', 'svc-prx-openresty', 'openresty') }}
       register: openresty_logs
       changed_when: false
       failed_when: false