diff --git a/ci/valkey/settings.env b/ci/valkey/settings.env
new file mode 100644
index 0000000..230ae70
--- /dev/null
+++ b/ci/valkey/settings.env
@@ -0,0 +1,4 @@
+# TEMPORARY, REVERTED IN THIS PR: the two negative experiments on this PR are
+# expected to time out, and 60s is enough to see them - the probe answers on the
+# first check and the init container fails before the server starts at all.
+INSTALL_TIMEOUT=60
diff --git a/ci/valkey/test-values.yaml b/ci/valkey/test-values.yaml
index 0db0486..81f265b 100644
--- a/ci/valkey/test-values.yaml
+++ b/ci/valkey/test-values.yaml
@@ -18,6 +18,17 @@ persistence:
   size: 128Mi
 
 valkey:
+  # TEMPORARY, REVERTED IN THIS PR: readiness authenticates with a deliberately
+  # wrong password while the server keeps the right one, to prove in the cluster
+  # that this probe CAN fail. Expected: startup passes, readiness fails with
+  # "NOAUTH Authentication required.", the pod stays 0/1 and the rollout wait
+  # times out.
+  readinessProbe:
+    exec:
+      command:
+        - /bin/sh
+        - -ec
+        - VALKEYCLI_AUTH=deliberately-wrong-password valkey-cli -e -p 6379 ping | grep -qx PONG
   # Small, so the whole thing stays inside the runner's budget. It still
   # exercises the derivation: 64Mi x limitFactor 3 = 192Mi limit, maxmemory =
   # 0.75 x that = 144Mi - and the render aborts if a configured maxmemory ever