fix(docker): use actual tmpfs for E2E test data

Replace misleading named volume with real tmpfs mount. E2E test data is
now truly ephemeral and fresh on every container start, with no state
leakage between test runs.

Fixes review feedback on PR #550.

.docker/compose/docker-compose.e2e.yml

@@ -31,16 +31,12 @@ services:
       - CHARON_CADDY_BINARY=caddy
       - CHARON_ACME_STAGING=true
       # FEATURE_CERBERUS_ENABLED deprecated - Cerberus enabled by default
-    volumes:
-      # Use tmpfs for E2E test data - fresh on every run
-      - e2e_data:/app/data
+    tmpfs:
+      # True tmpfs for E2E test data - fresh on every run, in-memory only
+      - /app/data:size=100M,mode=1755
     healthcheck:
       test: ["CMD-SHELL", "curl -fsS http://localhost:8080/api/v1/health || exit 1"]
       interval: 5s
       timeout: 5s
       retries: 10
       start_period: 10s
-
-volumes:
-  e2e_data:
-    driver: local