feat: adjustable deployment strategy (#118)

* feat: adjustable deployment strategy * chore: bump up chart version to 0.0.5 Signed-off-by: BrianPark314 <[email protected]>
vllm-project · Feb 13, 2025 · 748481e · 748481e
1 parent 82d2401
commit 748481e
Show file tree

Hide file tree

Showing 5 changed files with 34 additions and 5 deletions.
diff --git a/helm/Chart.yaml b/helm/Chart.yaml
@@ -15,7 +15,7 @@ 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: 0.0.4
+version: 0.0.5
 
 maintainers:
   - name: apostac
diff --git a/helm/templates/_helpers.tpl b/helm/templates/_helpers.tpl
@@ -31,14 +31,36 @@ Define container port name
 {{- end }}
 
 {{/*
-Define deployment strategy
+Define engine deployment strategy.
+If .Values.engineStrategy is defined, use it.
+Otherwise, fall back to the default rolling update strategy.
 */}}
-{{- define "chart.strategy" -}}
+{{- define "chart.engineStrategy" -}}
+{{- if .Values.servingEngineSpec.strategy }}
+{{ toYaml .Values.servingEngineSpec.strategy | nindent 2 }}
+{{- else }}
 strategy:
   rollingUpdate:
     maxSurge: 100%
     maxUnavailable: 0
 {{- end }}
+{{- end }}
+
+{{/*
+Define router deployment strategy.
+If .Values.routerStrategy is defined, use it.
+Otherwise, fall back to the default rolling update strategy.
+*/}}
+{{- define "chart.routerStrategy" -}}
+{{- if .Values.routerSpec.strategy }}
+{{ toYaml .Values.routerSpec.strategy | nindent 2 }}
+{{- else }}
+strategy:
+  rollingUpdate:
+    maxSurge: 100%
+    maxUnavailable: 0
+{{- end }}
+{{- end }}
 
 {{/*
 Define additional ports

diff --git a/helm/templates/deployment-router.yaml b/helm/templates/deployment-router.yaml
@@ -7,7 +7,8 @@ metadata:
   labels:
   {{- include "chart.routerLabels" . | nindent 4 }}
 spec:
-  replicas: 1
+  replicas: {{ .Values.routerSpec.replicaCount }}
+  {{- include "chart.routerStrategy" . | nindent 2 }}
   selector:
     matchLabels:
     {{- include "chart.routerLabels" . | nindent 6 }}

diff --git a/helm/templates/deployment-vllm-multi.yaml b/helm/templates/deployment-vllm-multi.yaml
@@ -9,7 +9,7 @@ metadata:
   {{- include "chart.engineLabels" . | nindent 4 }}
 spec:
   replicas: {{ $modelSpec.replicaCount }}
-  {{- include "chart.strategy" . | nindent 2 }}
+  {{- include "chart.engineStrategy" . | nindent 2 }}
   selector:
     matchLabels:
     {{- include "chart.engineLabels" . | nindent 6 }}

diff --git a/helm/values.yaml b/helm/values.yaml
@@ -92,6 +92,9 @@ servingEngineSpec:
   # -- Set other environment variables from config map
   configs: {}
 
+  # -- deployment strategy
+  strategy: {}
+
   # -- Readiness probe configuration
   startupProbe:
     # -- Number of seconds after the container has started before startup probe is initiated
@@ -171,6 +174,9 @@ routerSpec:
   # -- Window size in seconds to calculate the request statistics
   requestStatsWindow: 60
 
+  # -- deployment strategy
+  strategy: {}
+
   # -- router resource requests and limits
   resources:
     requests: