From 333fadad6cebe500ff68f702fcb9d9028ad2c6f2 Mon Sep 17 00:00:00 2001
From: yassin <yassin@berri.ai>
Date: Mon, 21 Sep 2026 06:25:43 +0000
Subject: [PATCH] fix(helm): render a fixed replicaCount on componentized
 deployments when HPA is disabled

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../litellm/templates/backend/deployment.yaml |  3 +
 .../litellm/templates/gateway/deployment.yaml |  3 +
 helm/litellm/templates/ui/deployment.yaml     |  3 +
 helm/litellm/tests/replica_count_tests.yaml   | 84 +++++++++++++++++++
 helm/litellm/values.yaml                      |  8 ++
 5 files changed, 101 insertions(+)
 create mode 100644 helm/litellm/tests/replica_count_tests.yaml

diff --git a/helm/litellm/templates/backend/deployment.yaml b/helm/litellm/templates/backend/deployment.yaml
index 0db2f0b3d43..6370f581d71 100644
--- a/helm/litellm/templates/backend/deployment.yaml
+++ b/helm/litellm/templates/backend/deployment.yaml
@@ -7,6 +7,9 @@ metadata:
     {{- include "litellm.commonLabels" . | nindent 4 }}
     app.kubernetes.io/component: backend
 spec:
+  {{- if not .Values.backend.hpa.enabled }}
+  replicas: {{ .Values.backend.replicaCount }}
+  {{- end }}
   {{- with .Values.backend.strategy }}
   strategy:
     {{- toYaml . | nindent 4 }}
diff --git a/helm/litellm/templates/gateway/deployment.yaml b/helm/litellm/templates/gateway/deployment.yaml
index c06cc9583a0..0c537b1bdeb 100644
--- a/helm/litellm/templates/gateway/deployment.yaml
+++ b/helm/litellm/templates/gateway/deployment.yaml
@@ -7,6 +7,9 @@ metadata:
     {{- include "litellm.commonLabels" . | nindent 4 }}
     app.kubernetes.io/component: gateway
 spec:
+  {{- if not .Values.gateway.hpa.enabled }}
+  replicas: {{ .Values.gateway.replicaCount }}
+  {{- end }}
   {{- with .Values.gateway.strategy }}
   strategy:
     {{- toYaml . | nindent 4 }}
diff --git a/helm/litellm/templates/ui/deployment.yaml b/helm/litellm/templates/ui/deployment.yaml
index b992b347bad..b794418b7e9 100644
--- a/helm/litellm/templates/ui/deployment.yaml
+++ b/helm/litellm/templates/ui/deployment.yaml
@@ -7,6 +7,9 @@ metadata:
     {{- include "litellm.commonLabels" . | nindent 4 }}
     app.kubernetes.io/component: ui
 spec:
+  {{- if not .Values.ui.hpa.enabled }}
+  replicas: {{ .Values.ui.replicaCount }}
+  {{- end }}
   {{- with .Values.ui.strategy }}
   strategy:
     {{- toYaml . | nindent 4 }}
diff --git a/helm/litellm/tests/replica_count_tests.yaml b/helm/litellm/tests/replica_count_tests.yaml
new file mode 100644
index 00000000000..e6a28689ff5
--- /dev/null
+++ b/helm/litellm/tests/replica_count_tests.yaml
@@ -0,0 +1,84 @@
+suite: test fixed replica count when HPA is disabled
+templates:
+  - gateway/deployment.yaml
+  - gateway/configmap.yaml
+  - backend/deployment.yaml
+  - ui/deployment.yaml
+values:
+  - ./values/required.yaml
+tests:
+  - it: gateway renders replicaCount into spec.replicas when its HPA is disabled
+    template: gateway/deployment.yaml
+    set:
+      gateway.hpa.enabled: false
+      gateway.replicaCount: 3
+    asserts:
+      - isKind:
+          of: Deployment
+      - equal:
+          path: spec.replicas
+          value: 3
+
+  - it: backend renders replicaCount into spec.replicas when its HPA is disabled
+    template: backend/deployment.yaml
+    set:
+      backend.hpa.enabled: false
+      backend.replicaCount: 2
+    asserts:
+      - equal:
+          path: spec.replicas
+          value: 2
+
+  - it: ui renders replicaCount into spec.replicas when its HPA is disabled
+    template: ui/deployment.yaml
+    set:
+      ui.hpa.enabled: false
+      ui.replicaCount: 2
+    asserts:
+      - equal:
+          path: spec.replicas
+          value: 2
+
+  - it: replicaCount 0 scales the gateway to zero instead of being treated as unset
+    template: gateway/deployment.yaml
+    set:
+      gateway.hpa.enabled: false
+      gateway.replicaCount: 0
+    asserts:
+      - equal:
+          path: spec.replicas
+          value: 0
+
+  - it: every component omits spec.replicas when its HPA is enabled, so the autoscaler owns the count
+    set:
+      gateway.hpa.enabled: true
+      gateway.replicaCount: 3
+      backend.hpa.enabled: true
+      backend.replicaCount: 3
+      ui.hpa.enabled: true
+      ui.replicaCount: 3
+    asserts:
+      - notExists:
+          path: spec.replicas
+        template: gateway/deployment.yaml
+      - notExists:
+          path: spec.replicas
+        template: backend/deployment.yaml
+      - notExists:
+          path: spec.replicas
+        template: ui/deployment.yaml
+
+  - it: a component with HPA disabled renders replicas while a sibling with HPA enabled does not
+    set:
+      gateway.hpa.enabled: false
+      gateway.replicaCount: 4
+      backend.hpa.enabled: true
+      backend.replicaCount: 4
+    asserts:
+      - equal:
+          path: spec.replicas
+          value: 4
+        template: gateway/deployment.yaml
+      - notExists:
+          path: spec.replicas
+        template: backend/deployment.yaml
diff --git a/helm/litellm/values.yaml b/helm/litellm/values.yaml
index 4ca54131d6a..d7836b03a34 100644
--- a/helm/litellm/values.yaml
+++ b/helm/litellm/values.yaml
@@ -397,6 +397,10 @@ gateway:
   #   failureThreshold: 30
   #   periodSeconds: 10
   startupProbe: {}
+  # Fixed pod count, rendered into the Deployment's spec.replicas only when
+  # hpa.enabled is false. With the HPA on, the autoscaler owns the count and
+  # this value is ignored.
+  replicaCount: 1
   hpa:
     enabled: true
     minReplicas: 1
@@ -524,6 +528,8 @@ backend:
   strategy: {}
   # Optional startupProbe; same shape as gateway.startupProbe. Empty by default.
   startupProbe: {}
+  # Same semantics as gateway.replicaCount.
+  replicaCount: 1
   hpa:
     enabled: true
     minReplicas: 1
@@ -590,6 +596,8 @@ ui:
   strategy: {}
   # Optional startupProbe; same shape as gateway.startupProbe. Empty by default.
   startupProbe: {}
+  # Same semantics as gateway.replicaCount.
+  replicaCount: 1
   hpa:
     enabled: false
     minReplicas: 1