Helm chart refactoring (#3797)

* initial commit for helm chart refactoring

* Continue refactoring helm. I was able to use helm to deploy all of the apps to a cluster in aws. The bottleneck was setting up PVC dynamic provisioning.

* use default storage class

* Fix linter errors

* Fix broken helm test

---------

Co-authored-by: jpb80 <jordan.buttkevitz@gmail.com>

.github/workflows/pr-helm-chart-testing.yml

@@ -21,10 +21,10 @@ jobs:
     - name: Set up Helm
       uses: azure/setup-helm@v4.2.0
       with:
-        version: v3.14.4
+        version: v3.17.0
       
     - name: Set up chart-testing
-      uses: helm/chart-testing-action@v2.6.1
+      uses: helm/chart-testing-action@v2.7.0
 
     # even though we specify chart-dirs in ct.yaml, it isn't used by ct for the list-changed command...
     - name: Run chart-testing (list-changed)
@@ -37,22 +37,6 @@ jobs:
           echo "changed=true" >> "$GITHUB_OUTPUT"
         fi
 
-#     rkuo: I don't think we need python?
-#     - name: Set up Python
-#       uses: actions/setup-python@v5
-#       with:
-#         python-version: '3.11'
-#         cache: 'pip'
-#         cache-dependency-path: |
-#           backend/requirements/default.txt
-#           backend/requirements/dev.txt
-#           backend/requirements/model_server.txt
-#     - run: |
-#         python -m pip install --upgrade pip
-#         pip install --retries 5 --timeout 30 -r backend/requirements/default.txt
-#         pip install --retries 5 --timeout 30 -r backend/requirements/dev.txt
-#         pip install --retries 5 --timeout 30 -r backend/requirements/model_server.txt
-
     # lint all charts if any changes were detected
     - name: Run chart-testing (lint)
       if: steps.list-changed.outputs.changed == 'true'
@@ -62,7 +46,7 @@ jobs:
 
     - name: Create kind cluster
       if: steps.list-changed.outputs.changed == 'true'
-      uses: helm/kind-action@v1.10.0
+      uses: helm/kind-action@v1.12.0
 
     - name: Run chart-testing (install)
       if: steps.list-changed.outputs.changed == 'true'

deployment/helm/charts/onyx/Chart.yaml

@@ -6,7 +6,7 @@ sources:
   - "https://github.com/onyx-dot-app/onyx"
 type: application
 version: 0.2.1
-appVersion: "latest"
+appVersion: latest
 annotations:
   category: Productivity
   licenses: MIT

deployment/helm/charts/onyx/templates/api-deployment.yaml

@@ -45,10 +45,10 @@ spec:
             - |
               alembic upgrade head &&
               echo "Starting Onyx Api Server" &&
-              uvicorn onyx.main:app --host 0.0.0.0 --port 8080
+              uvicorn onyx.main:app --host 0.0.0.0 --port {{ .Values.api.containerPorts.server }}
           ports:
             - name: api-server-port
-              containerPort: {{ .Values.api.service.port }}
+              containerPort: {{ .Values.api.containerPorts.server }}
               protocol: TCP
           resources:
             {{- toYaml .Values.api.resources | nindent 12 }}

deployment/helm/charts/onyx/templates/api-service.yaml

@@ -11,10 +11,10 @@ metadata:
 spec:
   type: {{ .Values.api.service.type }}
   ports:
-    - port: {{ .Values.api.service.port }}
-      targetPort: api-server-port
+    - port: {{ .Values.api.service.servicePort }}
+      targetPort: {{ .Values.api.service.targetPort }}
       protocol: TCP
-      name: api-server-port
+      name: {{ .Values.api.service.portName }}
   selector:
     {{- include "onyx-stack.selectorLabels" . | nindent 4 }}
     {{- if .Values.api.deploymentLabels }}

deployment/helm/charts/onyx/templates/indexing-model-deployment.yaml

@@ -5,7 +5,7 @@ metadata:
   labels:
     {{- include "onyx-stack.labels" . | nindent 4 }}
 spec:
-  replicas: 1
+  replicas: {{ .Values.indexCapability.replicaCount }}
   selector:
     matchLabels:
       {{- include "onyx-stack.selectorLabels" . | nindent 6 }}
@@ -25,12 +25,14 @@ spec:
         {{- end }}
     spec:
       containers:
-      - name: indexing-model-server
-        image: onyxdotapp/onyx-model-server:latest
-        imagePullPolicy: IfNotPresent
-        command: [ "uvicorn", "model_server.main:app", "--host", "0.0.0.0", "--port", "9000", "--limit-concurrency", "10" ]
+      - name: {{ .Values.indexCapability.name }}
+        image: "{{ .Values.indexCapability.image.repository }}:{{ .Values.indexCapability.image.tag | default .Chart.AppVersion }}"           
+        imagePullPolicy: {{ .Values.indexCapability.image.pullPolicy }}
+        command: [ "uvicorn", "model_server.main:app", "--host", "0.0.0.0", "--port", "{{ .Values.indexCapability.containerPorts.server }}", "--limit-concurrency", "{{ .Values.indexCapability.limitConcurrency }}" ]
         ports:
-        - containerPort: 9000
+        - name: model-server
+          containerPort: {{ .Values.indexCapability.containerPorts.server }}
+          protocol: TCP
         envFrom:
           - configMapRef:
               name: {{ .Values.config.envConfigMapName }}

deployment/helm/charts/onyx/templates/indexing-model-pvc.yaml

@@ -3,8 +3,9 @@ kind: PersistentVolumeClaim
 metadata:
   name: {{ .Values.indexCapability.indexingModelPVC.name }}
 spec:
+  storageClassName: {{ .Values.persistent.storageClassName }}
   accessModes:
     - {{ .Values.indexCapability.indexingModelPVC.accessMode | quote }}
   resources:
     requests:
-      storage: {{ .Values.indexCapability.indexingModelPVC.storage | quote }}
+      storage: {{ .Values.indexCapability.indexingModelPVC.storage | quote }}

deployment/helm/charts/onyx/templates/indexing-model-service.yaml

@@ -11,8 +11,8 @@ spec:
     {{- toYaml .Values.indexCapability.deploymentLabels | nindent 4 }}
     {{- end }}
   ports:
-    - name: {{ .Values.indexCapability.service.name }}
+    - name: {{ .Values.indexCapability.service.portName }}
       protocol: TCP
-      port: {{ .Values.indexCapability.service.port }}
-      targetPort: {{ .Values.indexCapability.service.port }}
-  type: {{ .Values.indexCapability.service.type }}
+      port: {{ .Values.indexCapability.service.servicePort  }}
+      targetPort: {{ .Values.indexCapability.service.targetPort }}
+  type: {{ .Values.indexCapability.service.type }}

deployment/helm/charts/onyx/templates/inference-model-deployment.yaml

@@ -3,14 +3,14 @@ kind: Deployment
 metadata:
   name: {{ include "onyx-stack.fullname" . }}-inference-model
   labels:
-    {{- range .Values.inferenceCapability.deployment.labels }}
+    {{- range .Values.inferenceCapability.labels }}
     {{ .key }}: {{ .value }}
     {{- end }}
 spec:
-  replicas: {{ .Values.inferenceCapability.deployment.replicas }}
+  replicas: {{ .Values.inferenceCapability.replicaCount }}
   selector:
     matchLabels:
-      {{- range .Values.inferenceCapability.deployment.labels }}
+      {{- range .Values.inferenceCapability.labels }}
       {{ .key }}: {{ .value }}
       {{- end }}
   template:
@@ -21,24 +21,26 @@ spec:
         {{- end }}
     spec:
       containers:
-      - name: {{ .Values.inferenceCapability.service.name }}
-        image: {{ .Values.inferenceCapability.deployment.image.repository }}:{{ .Values.inferenceCapability.deployment.image.tag }}
-        imagePullPolicy: {{ .Values.inferenceCapability.deployment.image.pullPolicy }}
-        command: {{ toYaml .Values.inferenceCapability.deployment.command | nindent 14 }}
+      - name: model-server-inference
+        image: "{{ .Values.inferenceCapability.image.repository }}:{{ .Values.inferenceCapability.image.tag | default .Chart.AppVersion }}"
+        imagePullPolicy: {{ .Values.inferenceCapability.image.pullPolicy }}
+        command: [ "uvicorn", "model_server.main:app", "--host", "0.0.0.0", "--port", "{{ .Values.inferenceCapability.containerPorts.server }}" ]
         ports:
-        - containerPort: {{ .Values.inferenceCapability.service.port }}
+        - name: model-server
+          containerPort: {{ .Values.inferenceCapability.containerPorts.server }}
+          protocol: TCP
         envFrom:
         - configMapRef:
             name: {{ .Values.config.envConfigMapName }}
         env:
           {{- include "onyx-stack.envSecrets" . | nindent 12}}
         volumeMounts:
-        {{- range .Values.inferenceCapability.deployment.volumeMounts }}
+        {{- range .Values.inferenceCapability.volumeMounts }}
         - name: {{ .name }}
           mountPath: {{ .mountPath }}
         {{- end }}
       volumes:
-      {{- range .Values.inferenceCapability.deployment.volumes }}
+      {{- range .Values.inferenceCapability.volumes }}
       - name: {{ .name }}
         persistentVolumeClaim:
           claimName: {{ .persistentVolumeClaim.claimName }}

deployment/helm/charts/onyx/templates/inference-model-pvc.yaml

@@ -3,6 +3,7 @@ kind: PersistentVolumeClaim
 metadata:
   name: {{ .Values.inferenceCapability.pvc.name }}
 spec:
+  storageClassName: {{ .Values.persistent.storageClassName }}
   accessModes: 
     {{- toYaml .Values.inferenceCapability.pvc.accessModes | nindent 4 }}
   resources:

deployment/helm/charts/onyx/templates/inference-model-service.yaml

@@ -5,11 +5,11 @@ metadata:
 spec:
   type: {{ .Values.inferenceCapability.service.type }}
   ports:
-    - port: {{ .Values.inferenceCapability.service.port }}
-      targetPort: {{ .Values.inferenceCapability.service.port }}
+    - port: {{ .Values.inferenceCapability.service.servicePort}}
+      targetPort: {{ .Values.inferenceCapability.service.targetPort }}
       protocol: TCP
-      name: {{ .Values.inferenceCapability.service.name }}
+      name: {{ .Values.inferenceCapability.service.portName }}
   selector:
-    {{- range .Values.inferenceCapability.deployment.labels }}
+    {{- range .Values.inferenceCapability.labels }}
     {{ .key }}: {{ .value }}
     {{- end }}

deployment/helm/charts/onyx/templates/nginx-conf.yaml

@@ -5,11 +5,11 @@ metadata:
 data:
   nginx.conf: |
     upstream api_server {
-        server {{ include "onyx-stack.fullname" . }}-api-service:{{ .Values.api.service.port }} fail_timeout=0;
+        server {{ include "onyx-stack.fullname" . }}-api-service:{{ .Values.api.service.servicePort }} fail_timeout=0;
     }
 
     upstream web_server {
-        server {{ include "onyx-stack.fullname" . }}-webserver:{{ .Values.webserver.service.port }} fail_timeout=0;
+        server {{ include "onyx-stack.fullname" . }}-webserver:{{ .Values.webserver.service.servicePort }} fail_timeout=0;
     }
 
     server {

deployment/helm/charts/onyx/templates/tests/test-connection.yaml

@@ -11,5 +11,5 @@ spec:
     - name: wget
       image: busybox
       command: ['wget']
-      args: ['{{ include "onyx-stack.fullname" . }}-webserver:{{ .Values.webserver.service.port }}']
+      args: ['{{ include "onyx-stack.fullname" . }}-webserver:{{ .Values.webserver.service.servicePort }}']
   restartPolicy: Never

deployment/helm/charts/onyx/templates/webserver-deployment.yaml

@@ -41,7 +41,7 @@ spec:
           imagePullPolicy: {{ .Values.webserver.image.pullPolicy }}
           ports:
             - name: http
-              containerPort: {{ .Values.webserver.service.port }}
+              containerPort: {{ .Values.webserver.containerPorts.server }}
               protocol: TCP
           resources:
             {{- toYaml .Values.webserver.resources | nindent 12 }}

deployment/helm/charts/onyx/templates/webserver-service.yaml

@@ -10,8 +10,8 @@ metadata:
 spec:
   type: {{ .Values.webserver.service.type }}
   ports:
-    - port: {{ .Values.webserver.service.port }}
-      targetPort: http
+    - port: {{ .Values.webserver.service.servicePort }}
+      targetPort: {{ .Values.webserver.service.targetPort }}
       protocol: TCP
       name: http
   selector:

deployment/helm/charts/onyx/values.yaml

@@ -2,62 +2,73 @@
 # This is a YAML-formatted file.
 # Declare variables to be passed into your templates.
 
+postgresql:
+  primary:
+    persistence:
+      size: 5Gi
+  enabled: true
+  auth:
+    existingSecret: onyx-secrets
+    secretKeys:
+      # overwriting as postgres typically expects 'postgres-password'
+      adminPasswordKey: postgres_password
 imagePullSecrets: []
 nameOverride: ""
 fullnameOverride: ""
 
+persistent:
+  storageClassName: ""
+
 inferenceCapability:
   service:
-    name: inference-model-server-service
+    portName: modelserver
     type: ClusterIP
-    port: 9000
+    servicePort: 9000
+    targetPort: 9000
   pvc:
     name: inference-model-pvc
     accessModes:
       - ReadWriteOnce
     storage: 3Gi
-  deployment:
-    name: inference-model-server-deployment
-    replicas: 1
-    labels:
-      - key: app
-        value: inference-model-server
-    image:
-      repository: onyxdotapp/onyx-model-server
-      tag: latest
-      pullPolicy: IfNotPresent
-    command:
-      [
-        "uvicorn",
-        "model_server.main:app",
-        "--host",
-        "0.0.0.0",
-        "--port",
-        "9000",
-      ]
-    port: 9000
-    volumeMounts:
-      - name: inference-model-storage
-        mountPath: /root/.cache
-    volumes:
-      - name: inference-model-storage
-        persistentVolumeClaim:
-          claimName: inference-model-pvc
+  name: inference-model-server
+  replicaCount: 1
+  labels:
+    - key: app
+      value: inference-model-server
+  image:
+    repository: onyxdotapp/onyx-model-server
+    # Overrides the image tag whose default is the chart appVersion.
+    tag: ""
+    pullPolicy: IfNotPresent
+  containerPorts:
+    server: 9000
+  volumeMounts:
+    - name: inference-model-storage
+      mountPath: /root/.cache
+  volumes:
+    - name: inference-model-storage
+      persistentVolumeClaim:
+        claimName: inference-model-pvc
   podLabels:
     - key: app
       value: inference-model-server
 
 indexCapability:
   service:
+    portName: modelserver
     type: ClusterIP
-    port: 9000
-    name: indexing-model-server-port
+    servicePort: 9000
+    targetPort: 9000
+  replicaCount: 1
+  name: indexing-model-server
   deploymentLabels:
     app: indexing-model-server
   podLabels:
     app: indexing-model-server
   indexingOnly: "True"
   podAnnotations: {}
+  containerPorts:
+    server: 9000
   volumeMounts:
     - name: indexing-model-storage
       mountPath: /root/.cache
@@ -69,7 +80,12 @@ indexCapability:
     name: indexing-model-storage
     accessMode: "ReadWriteOnce"
     storage: "3Gi"
-
+  image:
+    repository: onyxdotapp/onyx-model-server
+    # Overrides the image tag whose default is the chart appVersion.
+    tag: ""
+    pullPolicy: IfNotPresent
+  limitConcurrency: 10
 config:
   envConfigMapName: env-configmap
 
@@ -84,16 +100,6 @@ serviceAccount:
   # If not set and create is true, a name is generated using the fullname template
   name: ""
 
-postgresql:
-  primary:
-    persistence:
-      size: 5Gi
-  enabled: true
-  auth:
-    existingSecret: onyx-secrets
-    secretKeys:
-      adminPasswordKey: postgres_password # overwriting as postgres typically expects 'postgres-password'
-
 nginx:
   containerPorts:
     http: 1024
@@ -135,9 +141,13 @@ webserver:
     # runAsNonRoot: true
     # runAsUser: 1000
 
+  containerPorts:
+    server: 3000
+
   service:
     type: ClusterIP
-    port: 3000
+    servicePort: 3000
+    targetPort: http
 
   resources: {}
   # We usually recommend not to specify default resources and to leave this as a conscious
@@ -156,7 +166,7 @@ webserver:
     minReplicas: 1
     maxReplicas: 100
     targetCPUUtilizationPercentage: 80
-    # targetMemoryUtilizationPercentage: 80
+    targetMemoryUtilizationPercentage: 80
 
   # Additional volumes on the output Deployment definition.
   volumes: []
@@ -189,6 +199,9 @@ api:
     scope: onyx-backend
     app: api-server
 
+  containerPorts:
+    server: 8080
+
   podSecurityContext:
     {}
     # fsGroup: 2000
@@ -204,7 +217,9 @@ api:
 
   service:
     type: ClusterIP
-    port: 8080
+    servicePort: 8080
+    targetPort: api-server-port
+    portName: api-server-port
 
   resources: {}
   # We usually recommend not to specify default resources and to leave this as a conscious
@@ -223,7 +238,7 @@ api:
     minReplicas: 1
     maxReplicas: 100
     targetCPUUtilizationPercentage: 80
-    # targetMemoryUtilizationPercentage: 80
+    targetMemoryUtilizationPercentage: 80
 
   # Additional volumes on the output Deployment definition.
   volumes: []
@@ -247,7 +262,7 @@ background:
     repository: onyxdotapp/onyx-backend
     pullPolicy: IfNotPresent
     # Overrides the image tag whose default is the chart appVersion.
-    tag: latest
+    tag: ""
   podAnnotations: {}
   podLabels:
     scope: onyx-backend
@@ -284,7 +299,7 @@ background:
     minReplicas: 1
     maxReplicas: 100
     targetCPUUtilizationPercentage: 80
-    # targetMemoryUtilizationPercentage: 80
+    targetMemoryUtilizationPercentage: 80
 
   # Additional volumes on the output Deployment definition.
   volumes: []
@@ -303,6 +318,16 @@ background:
   tolerations: []
 
 vespa:
+  volumeClaimTemplates:
+    - metadata:
+        name: vespa-storage
+      spec:
+        accessModes:
+          - ReadWriteOnce
+        storageClassName: ""
+        resources:
+          requests:
+            storage: 1Gi
   enabled: true
   replicaCount: 1
   image:
@@ -377,19 +402,11 @@ redis:
 #  #    hosts:
 #  #      - chart-example.local
 
-persistence:
-  vespa:
-    enabled: true
-    existingClaim: ""
-    storageClassName: ""
-    accessModes:
-      - ReadWriteOnce
-    size: 5Gi
 
 auth:
-  # for storing smtp, oauth, slack, and other secrets
+  # existingSecret onyx-secret for storing smtp, oauth, slack, and other secrets
   # keys are lowercased version of env vars (e.g. SMTP_USER -> smtp_user)
-  existingSecret: "" # onyx-secrets
+  existingSecret: ""
   # optionally override the secret keys to reference in the secret
   # this is used to populate the env vars in individual deployments
   # the values here reference the keys in secrets below
@@ -413,14 +430,22 @@ auth:
     redis_password: "password"
 
 configMap:
-  AUTH_TYPE: "disabled" # Change this for production uses unless Onyx is only accessible behind VPN
-  SESSION_EXPIRE_TIME_SECONDS: "86400" # 1 Day Default
-  VALID_EMAIL_DOMAINS: "" # Can be something like onyx.app, as an extra double-check
-  SMTP_SERVER: "" # For sending verification emails, if unspecified then defaults to 'smtp.gmail.com'
-  SMTP_PORT: "" # For sending verification emails, if unspecified then defaults to '587'
-  SMTP_USER: "" # 'your-email@company.com'
-  # SMTP_PASS: ""  # 'your-gmail-password'
-  EMAIL_FROM: "" # 'your-email@company.com' SMTP_USER missing used instead
+  # Change this for production uses unless Onyx is only accessible behind VPN
+  AUTH_TYPE: "disabled"
+  # 1 Day Default
+  SESSION_EXPIRE_TIME_SECONDS: "86400"
+  # Can be something like onyx.app, as an extra double-check
+  VALID_EMAIL_DOMAINS: ""
+  # For sending verification emails, if unspecified then defaults to 'smtp.gmail.com'
+  SMTP_SERVER: ""
+  # For sending verification emails, if unspecified then defaults to '587'
+  SMTP_PORT: ""
+# 'your-email@company.com'
+  SMTP_USER: ""
+  # 'your-gmail-password'
+  # SMTP_PASS: ""
+  # 'your-email@company.com' SMTP_USER missing used instead
+  EMAIL_FROM: ""
   # Gen AI Settings
   GEN_AI_MAX_TOKENS: ""
   QA_TIMEOUT: "60"
@@ -462,7 +487,7 @@ configMap:
   DANSWER_BOT_DISABLE_DOCS_ONLY_ANSWER: ""
   DANSWER_BOT_DISPLAY_ERROR_MSGS: ""
   DANSWER_BOT_RESPOND_EVERY_CHANNEL: ""
-  DANSWER_BOT_DISABLE_COT: "" # Currently unused
+  DANSWER_BOT_DISABLE_COT: ""
   NOTIFY_SLACKBOT_NO_ANSWER: ""
   # Logging
   # Optional Telemetry, please keep it on (nothing sensitive is collected)? <3
@@ -473,7 +498,8 @@ configMap:
   LOG_DANSWER_MODEL_INTERACTIONS: ""
   LOG_VESPA_TIMING_INFORMATION: ""
   # Shared or Non-backend Related
-  WEB_DOMAIN: "http://localhost:3000" # for web server and api server
-  DOMAIN: "localhost" # for nginx
+  WEB_DOMAIN: "http://localhost:3000"
+  # DOMAIN used by nginx
+  DOMAIN: "localhost"
   # Chat Configs
   HARD_DELETE_CHATS: ""