diff --git a/helm/Chart.yaml b/helm/Chart.yaml
index 87ffcd5..67c4b8c 100644
--- a/helm/Chart.yaml
+++ b/helm/Chart.yaml
@@ -2,8 +2,8 @@ apiVersion: v2
 name: socket-firewall
 description: Socket.dev Registry Firewall - Block vulnerable packages before they reach your cluster
 type: application
-version: 0.4.1
-appVersion: "1.1.335"
+version: 0.5.0
+appVersion: "1.1.341"
 keywords:
   - security
   - supply-chain
diff --git a/helm/README.md b/helm/README.md
index b19cddc..c23019b 100644
--- a/helm/README.md
+++ b/helm/README.md
@@ -122,6 +122,7 @@ registries:
 | `ingress.className` | Ingress class (nginx, alb, traefik) | `""` |
 | `autoscaling.enabled` | Enable HorizontalPodAutoscaler | `false` |
 | `podDisruptionBudget.enabled` | Keep pods available during node maintenance | `true` |
+| `topologySpreadConstraints` | Evenly spread replicas across zones/nodes | `[]` |
 | `extraContainers` | Sidecar containers (auth proxies, log collectors) | `[]` |
 | `resources.limits.cpu` | CPU limit | `4` |
 | `resources.limits.memory` | Memory limit | `8Gi` |
@@ -492,6 +493,35 @@ kubectl get hpa socket-firewall
 kubectl describe hpa socket-firewall
 ```
 
+## Spreading Replicas Across Zones
+
+When you run more than one replica (via `replicaCount` or autoscaling), use
+`topologySpreadConstraints` to distribute pods evenly across availability zones
+(or nodes) so a single zone/node failure can't take down a disproportionate share
+of the fleet. This is preferred over soft pod anti-affinity, which the scheduler is
+free to ignore and can pile replicas into one zone.
+
+```yaml
+replicaCount: 3
+topologySpreadConstraints:
+  - maxSkew: 1
+    topologyKey: topology.kubernetes.io/zone
+    whenUnsatisfiable: ScheduleAnyway   # best-effort even spread; never blocks scheduling
+    labelSelector:
+      matchLabels:
+        app.kubernetes.io/name: socket-firewall
+```
+
+- `maxSkew: 1` keeps zones within one pod of each other.
+- `whenUnsatisfiable: ScheduleAnyway` is a soft guarantee (recommended). Use
+  `DoNotSchedule` for a hard guarantee — but be aware pods can stay `Pending` if a
+  zone is full or you have fewer zones than replicas.
+- Pin spreading to a specific revision by adding `matchLabelKeys: [pod-template-hash]`
+  (requires Kubernetes 1.27+, satisfied by all currently-supported EKS and GKE versions).
+
+**Compatibility:** `topologySpreadConstraints` is GA since Kubernetes 1.19, so it works
+on every currently-supported cluster. The chart sets no `kubeVersion` floor.
+
 ## Using an Existing Secret for API Token
 
 ```bash
diff --git a/helm/templates/deployment.yaml b/helm/templates/deployment.yaml
index 0b510d3..e4c639b 100644
--- a/helm/templates/deployment.yaml
+++ b/helm/templates/deployment.yaml
@@ -369,6 +369,10 @@ spec:
       affinity:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.tolerations }}
       tolerations:
         {{- toYaml . | nindent 8 }}
diff --git a/helm/values.yaml b/helm/values.yaml
index 92f7275..5fcc5e1 100644
--- a/helm/values.yaml
+++ b/helm/values.yaml
@@ -5,7 +5,7 @@ image:
   repository: socketdev/socket-registry-firewall
   # Defaults to the latest published image so deployments pick up new releases
   # without a chart bump. Pin a version for reproducibility, e.g. tag: "1.1.327".
-  tag: "1.1.335"
+  tag: "1.1.341"
   # Always (not IfNotPresent) so "latest" actually refreshes on pod restart.
   pullPolicy: Always
 
@@ -432,6 +432,22 @@ tolerations: []
 # Affinity
 affinity: {}
 
+# Topology spread constraints (evenly distribute replicas across failure domains
+# such as zones or nodes). Preferred over soft pod anti-affinity for even spreading.
+# Empty by default; only takes effect with replicaCount > 1 or autoscaling enabled.
+# The base field is GA since Kubernetes 1.19 (universally available). The optional
+# matchLabelKeys field needs Kubernetes 1.27+ (satisfied by all current EKS/GKE versions).
+topologySpreadConstraints: []
+  # - maxSkew: 1
+  #   topologyKey: topology.kubernetes.io/zone
+  #   whenUnsatisfiable: ScheduleAnyway   # use DoNotSchedule for a hard guarantee
+  #   labelSelector:
+  #     matchLabels:
+  #       app.kubernetes.io/name: socket-firewall
+  #   # matchLabelKeys requires Kubernetes 1.27+
+  #   # matchLabelKeys:
+  #   #   - pod-template-hash
+
 # Extra containers to run alongside the firewall (e.g., auth sidecars, log collectors)
 extraContainers: []
 # - name: auth-sidecar