diff --git a/argocd/manifests/grafana/deployment.yaml b/argocd/manifests/grafana/deployment.yaml
index 0aad9b3..cbba267 100644
--- a/argocd/manifests/grafana/deployment.yaml
+++ b/argocd/manifests/grafana/deployment.yaml
@@ -14,7 +14,9 @@ spec:
       app.kubernetes.io/name: grafana
       app.kubernetes.io/instance: grafana
   strategy:
-    type: RollingUpdate
+    # RWO PVC for SQLite + Bleve index — RollingUpdate spawns the new pod
+    # before the old one terminates, and it crashloops on the index lock.
+    type: Recreate
   template:
     metadata:
       labels:
diff --git a/docs/changelog.d/+grafana-recreate-strategy.infra.md b/docs/changelog.d/+grafana-recreate-strategy.infra.md
new file mode 100644
index 0000000..3662e10
--- /dev/null
+++ b/docs/changelog.d/+grafana-recreate-strategy.infra.md
@@ -0,0 +1 @@
+Switched Grafana's deployment strategy from `RollingUpdate` to `Recreate`. With an RWO PVC holding the SQLite database and Bleve search index, `RollingUpdate` reliably crashloops the new pod on the index lock until rollout timeout. `Recreate` terminates the old pod first so the new one acquires the lock cleanly.