diff --git a/charts/service/Chart.yaml b/charts/service/Chart.yaml
index 53e12b3..225fdd3 100644
--- a/charts/service/Chart.yaml
+++ b/charts/service/Chart.yaml
@@ -20,7 +20,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.5.3
+version: 0.5.4
 
 home: https://github.com/Breadfast/helm-chart
 sources:
diff --git a/charts/service/README.md b/charts/service/README.md
index 06676c0..5932a8f 100644
--- a/charts/service/README.md
+++ b/charts/service/README.md
@@ -1,6 +1,6 @@
 # service
 
-![Version: 0.5.3](https://img.shields.io/badge/Version-0.5.3-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square)
+![Version: 0.5.4](https://img.shields.io/badge/Version-0.5.4-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square)
 
 A Helm chart for Kubernetes.
 
@@ -122,6 +122,7 @@ networkPolicy.internetOnly.ipBlock.except when they are outside the defaults.
 | podAntiAffinity | object | `{}` |  |
 | podLabels | object | `{}` |  |
 | podSecurityContext | object | `{}` |  |
+| priorityClassName | string | `""` | Name of an existing PriorityClass. Empty means cluster default (priority 0). |
 | readinessProbe | object | `{}` |  |
 | replicaCount | int | `1` |  |
 | resources | object | `{}` |  |
@@ -137,6 +138,7 @@ networkPolicy.internetOnly.ipBlock.except when they are outside the defaults.
 | serviceAccount.create | bool | `false` | If true, creates service account |
 | serviceAccount.name | string | `""` |  |
 | startupProbe | object | `{}` |  |
+| topologySpreadConstraints | list | `[]` | Spread replicas across nodes or zones. See the comment above for how this differs from podAntiAffinity. |
 | vaultAgent | bool | `{"enabled":false}` | If true, It will inject Vault Agent to get secrets from Vault |
 | virtualService.enabled | bool | `false` |  |
 | virtualService.hosts | list | `[]` |  |
diff --git a/charts/service/templates/argo-rollouts.yaml b/charts/service/templates/argo-rollouts.yaml
index f1aceb8..0e077e7 100644
--- a/charts/service/templates/argo-rollouts.yaml
+++ b/charts/service/templates/argo-rollouts.yaml
@@ -114,10 +114,17 @@ spec:
         podAntiAffinity:
           {{- toYaml . | nindent 10 }}
         {{- end }}
+      {{- with .Values.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.nodeSelectorLabels }}
       nodeSelector:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.priorityClassName }}
+      priorityClassName: {{ . }}
+      {{- end }}
       {{- with .Values.hostAliases }}
       hostAliases:
         {{- toYaml . | nindent 8 }}
diff --git a/charts/service/templates/deployment.yaml b/charts/service/templates/deployment.yaml
index b59762f..9e28310 100644
--- a/charts/service/templates/deployment.yaml
+++ b/charts/service/templates/deployment.yaml
@@ -55,10 +55,17 @@ spec:
         podAntiAffinity:
           {{- toYaml . | nindent 10 }}
         {{- end }}
+      {{- with .Values.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.nodeSelectorLabels }}
       nodeSelector:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.priorityClassName }}
+      priorityClassName: {{ . }}
+      {{- end }}
       {{- with .Values.hostAliases }}
       hostAliases:
         {{- toYaml . | nindent 8 }}
diff --git a/charts/service/values.yaml b/charts/service/values.yaml
index 560be53..3ef374d 100644
--- a/charts/service/values.yaml
+++ b/charts/service/values.yaml
@@ -253,6 +253,48 @@ entrypointOverride:
 podAffinity: {}
 podAntiAffinity: {}
 
+# Prefer this over podAntiAffinity when spreading replicas across nodes.
+#
+# Skew is evaluated across the topology domains actually eligible for the pod, so
+# `maxSkew: 1` lands a 5-replica Deployment 3/2 across two usable nodes. A
+# `required` podAntiAffinity instead demands one node per replica and leaves the
+# rest Pending, which on a saturated pool forces continuous autoscaler scale-up.
+#
+# whenUnsatisfiable: ScheduleAnyway keeps it advisory but still skew-aware;
+# DoNotSchedule makes it a hard guarantee within the maxSkew tolerance.
+#
+# Example:
+#   topologySpreadConstraints:
+#     - maxSkew: 1
+#       topologyKey: kubernetes.io/hostname
+#       whenUnsatisfiable: DoNotSchedule
+#       labelSelector:
+#         matchLabels:
+#           app.kubernetes.io/name: my-service
+#
+# Docs: https://kubernetes.io/docs/concepts/scheduling-eviction/topology-spread-constraints/
+# -- (list) Spread replicas across nodes or zones. See the comment above for how this differs from podAntiAffinity.
+topologySpreadConstraints: []
+
+# Assign a PriorityClass so the scheduler knows which pods matter when the cluster
+# is short on room. Unset means priority 0 — the same as every workload today,
+# since no PriorityClass in this cluster sets globalDefault.
+#
+# The PriorityClass must already exist in the cluster; `kubectl get priorityclasses`
+# lists them. Raising a workload's priority lets the scheduler PREEMPT (evict)
+# lower-priority pods to make room for it, so treat a value above 0 as a decision
+# about which of your OWN services get displaced, not a free win. Use
+# preemptionPolicy: Never on the PriorityClass to get ordering without eviction.
+#
+# WATCH OUT: a name that does not exist fails LATE and quietly. The Deployment is
+# accepted by the API (verified against the live 1.35 apiserver), so ArgoCD reports
+# Synced/Healthy, but every Pod is then rejected with
+#   forbidden: no PriorityClass with name <x> was found
+# so the new ReplicaSet sits at 0 replicas while maxUnavailable keeps the old pods
+# serving. Confirm the name with `kubectl get priorityclasses` before setting it.
+# -- (string) Name of an existing PriorityClass. Empty means cluster default (priority 0).
+priorityClassName: ""
+
 # -- (bool) If true, Creates CronJob resource
 cronJob:
   create: false