diff --git a/tests/ghost/compose.ccci.yml b/tests/ghost/compose.ccci.yml
index 9995168..040b156 100644
--- a/tests/ghost/compose.ccci.yml
+++ b/tests/ghost/compose.ccci.yml
@@ -29,8 +29,17 @@
 # bites BOTH base and head (the published recipe ships db start_period 1m everywhere), so the overlay
 # applies on both (persists untracked across the head checkout) — a recipe-PR candidate too.
 # Grace-only; masks no defect; weakens no test.
+#
+# The app also needs a LONGER restart delay during the base→head crossover. On current Ghost heads the
+# upgrade concurrently replaces mysql 8.0 with mysql 8.4; swarm starts a new app task before the new
+# db service name is resolvable/ready, so Ghost exits immediately with `ENOTFOUND <stack>_db` and the
+# rolling update pauses before the next retry. A 15s delay makes the retried app start after mysql is
+# back, which preserves the real upgrade/assertion while removing a transient service-discovery race.
 services:
   app:
+    deploy:
+      restart_policy:
+        delay: 15s
     healthcheck:
       start_period: 15m
   db: