n8n: add bounded retries and timeout to WF1's Fleet Ops callback

Vehicle Return Orchestration had no explicit timeout and Retry On Fail
disabled on its outbound HTTP call, a gap against the acceptance
checklist's timeouts/bounded-retries requirement. Fixed live (3 tries,
1000ms wait, 15s timeout, matching WF2's existing convention) and
synced the repo definition + manifest checksum.

n8n/workflows/fleet-ops-vehicle-return.json

@@ -51,7 +51,9 @@
         "contentType": "raw",
         "rawContentType": "application/json",
         "body": "={{JSON.stringify($json)}}",
-        "options": {}
+        "options": {
+          "timeout": 15000
+        }
       },
       "id": "callback-node",
       "name": "Record follow-up",
@@ -62,8 +64,10 @@
         "httpHeaderAuth": {
           "name": "Fleet Ops Service Token"
         }
-      }
-    },
+      },
+      "retryOnFail": true,
+      "maxTries": 3,
+      "waitBetweenTries": 1000
     {
       "parameters": {
         "respondWith": "json",