Merge pull request #190 from jyje/issue-184/add-milvus

feat(charts/open-webui): add the Milvus chart for RAG

.github/workflows/helm-test-open-webui.yml

@@ -30,6 +30,7 @@ jobs:
           helm repo add open-webui https://helm.openwebui.com/
           helm repo add tika https://apache.jfrog.io/artifactory/tika/
           helm repo add redis https://charts.bitnami.com/bitnami
+          helm repo add milvus https://zilliztech.github.io/milvus-helm
       - name: Build open-webui Helm dependencies
         run: |
           helm dependency build ./charts/open-webui

charts/open-webui/Chart.lock

@@ -10,6 +10,9 @@ dependencies:
   version: 2.9.0
 - name: redis
   repository: https://charts.bitnami.com/bitnami
-  version: 20.9.0
-digest: sha256:d94d9a3b5d723c49a041c006020f0e935e9b61c1070c40310c0974386114b718
-generated: "2025-02-26T06:40:17.918812+09:00"
+  version: 20.10.0
+- name: milvus
+  repository: https://zilliztech.github.io/milvus-helm
+  version: 4.2.40
+digest: sha256:b9597e9cf5f89874a3c345562085ac3ba972b98ccf1b5e64425de0c17ec359f5
+generated: "2025-02-28T22:43:15.040921+09:00"

charts/open-webui/Chart.yaml

@@ -1,6 +1,6 @@
 apiVersion: v2
 name: open-webui
-version: 5.20.0
+version: 5.21.0
 appVersion: 0.5.16
 home: https://www.openwebui.com/
 icon: >-
@@ -43,3 +43,7 @@ dependencies:
     version: '>=20.6.2'
     alias: redis-cluster
     condition: redis-cluster.enabled
+  - name: milvus
+    repository: https://zilliztech.github.io/milvus-helm
+    version: '>=4.2.40'
+    condition: milvus.enabled

charts/open-webui/README.md

@@ -1,6 +1,6 @@
 # open-webui
 
-![Version: 5.20.0](https://img.shields.io/badge/Version-5.20.0-informational?style=flat-square) ![AppVersion: 0.5.16](https://img.shields.io/badge/AppVersion-0.5.16-informational?style=flat-square)
+![Version: 5.21.0](https://img.shields.io/badge/Version-5.21.0-informational?style=flat-square) ![AppVersion: 0.5.16](https://img.shields.io/badge/AppVersion-0.5.16-informational?style=flat-square)
 
 Open WebUI: A User-Friendly Web Interface for Chat Interactions 👋
 
@@ -37,6 +37,7 @@ helm upgrade --install open-webui open-webui/open-webui
 | https://charts.bitnami.com/bitnami | redis-cluster(redis) | >=20.6.2 |
 | https://helm.openwebui.com | pipelines | >=0.0.1 |
 | https://otwld.github.io/ollama-helm/ | ollama | >=0.24.0 |
+| https://zilliztech.github.io/milvus-helm | milvus | >=4.2.40 |
 
 ## Values
 
@@ -64,6 +65,11 @@ helm upgrade --install open-webui open-webui/open-webui
 | managedCertificate.domains[0] | string | `"chat.example.com"` |  |
 | managedCertificate.enabled | bool | `false` |  |
 | managedCertificate.name | string | `"mydomain-chat-cert"` |  |
+| milvus.db | string | `"default"` | Active Milvus database for RAG with env `MILVUS_DB` ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_db |
+| milvus.enabled | bool | `false` | Enable Milvus installation. Deploys a Milvus cluster/standalone with subchart 'milvus' from zilliztech ref: https://github.com/zilliztech/milvus-helm/tree/master/charts/milvus |
+| milvus.fullnameOverride | string | `"open-webui-milvus"` | Milvus fullname override (recommended to be 'open-webui-milvus') - In this case, the Milvus uri will be 'http://[username:password@]open-webui-milvus:19530' |
+| milvus.token | object | `{}` | Active Milvus token for RAG with env `MILVUS_TOKEN` ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_token |
+| milvus.uri | string | `"http://open-webui-milvus:19530"` | Active Milvus URI for RAG with env `MILVUS_URI`. If there is credentials in the uri, it will be used to connect to the Milvus server. ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_uri |
 | nameOverride | string | `""` |  |
 | namespaceOverride | string | `""` |  |
 | nodeSelector | object | `{}` | Node labels for pod assignment. |
@@ -87,6 +93,10 @@ helm upgrade --install open-webui open-webui/open-webui
 | podAnnotations | object | `{}` |  |
 | podLabels | object | `{}` |  |
 | podSecurityContext | object | `{}` | Configure pod security context ref: <https://kubernetes.io/docs/tasks/configure-pod-container/security-context/#set-the-security-context-for-a-containe> |
+| rag.embeddingEngine | string | `""` | Embedding engine to use for RAG with env `RAG_EMBEDDING_ENGINE`: ""(empty), "ollama", "openai" ref: https://docs.openwebui.com/getting-started/env-configuration#rag_embedding_engine |
+| rag.embeddingModel | string | `""` | Embedding model to use for RAG with env `RAG_EMBEDDING_MODEL` ref: https://docs.openwebui.com/getting-started/env-configuration#rag_embedding_model |
+| rag.enabled | bool | `false` | Enable RAG ref: https://docs.openwebui.com/getting-started/env-configuration#retrieval-augmented-generation-rag |
+| rag.vectorDB | string | `""` | Vector database configuration ref: https://docs.openwebui.com/getting-started/env-configuration#vector_db |
 | readinessProbe | object | `{}` | Probe for readiness of the Open WebUI container ref: <https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes> |
 | redis-cluster | object | `{"auth":{"enabled":false},"enabled":false,"fullnameOverride":"open-webui-redis","replica":{"replicaCount":3}}` | Deploys a Redis cluster with subchart 'redis' from bitnami |
 | redis-cluster.auth | object | `{"enabled":false}` | Redis Authentication |

charts/open-webui/templates/workload-manager.yaml

@@ -141,6 +141,28 @@ spec:
         - name: "ENABLE_OPENAI_API"
           value: "False"
         {{- end }}
+        {{- if .Values.rag.enabled }}
+        - name: "VECTOR_DB"
+          value: {{ .Values.rag.vectorDB | default "croma" | quote }}
+        {{- if and .Values.rag.enabled .Values.rag.embeddingEngine }}
+        - name: "RAG_EMBEDDING_ENGINE"
+          value: {{ .Values.rag.embeddingEngine | quote }}
+        {{- end }}
+        {{- if and .Values.rag.enabled .Values.rag.embeddingModel }}
+        - name: "RAG_EMBEDDING_MODEL"
+          value: {{ .Values.rag.embeddingModel | quote }}
+        {{- end }}
+        {{- end }}
+        {{- if .Values.milvus.enabled }}
+        - name: "MILVUS_URI"
+          value: {{ .Values.milvus.uri | default "${DATA_DIR}/vector_db/milvus.db" | quote }}
+        - name: "MILVUS_DB"
+          value: {{ .Values.milvus.db | default "default" | quote }}
+        {{- if and .Values.milvus.enabled .Values.milvus.token }}
+        - name: "MILVUS_TOKEN"
+          value: {{ .Values.milvus.token | quote }}
+        {{- end }}
+        {{- end }}
         {{- if .Values.extraEnvVars }}
           {{- toYaml .Values.extraEnvVars | nindent 8 }}
         {{- end }}

charts/open-webui/values-rag-milvus.yaml

@@ -0,0 +1,53 @@
+rag:
+  # -- Enable RAG
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#retrieval-augmented-generation-rag
+  enabled: true
+  vectorDB: milvus
+  embeddingEngine: ""
+  embeddingModel: ""
+
+milvus:
+  # -- Enable Milvus installation. Deploys a Milvus cluster/standalone with subchart 'milvus' from zilliztech
+  # ref: https://github.com/zilliztech/milvus-helm/tree/master/charts/milvus
+  enabled: true
+  uri: "http://open-webui-milvus:19530"
+  db: default
+  token: {}
+  cluster:
+    enabled: false # This means that the Milvus runs with standalone mode
+  minio:
+    enabled: true
+    resources:
+      requests:
+        memory: 50Mi
+    persistence:
+      enabled: true
+      size: 1Gi
+  etcd:
+    enabled: true
+  pulsar:
+    enabled: false
+  pulsarv3:
+    enabled: false
+  kafka:
+    enabled: false
+  externalS3:
+    enabled: false
+  externalEtcd:
+    enabled: false
+
+livenessProbe:
+  httpGet:
+    path: /health
+    port: http
+readinessProbe:
+  httpGet:
+    path: /health/db
+    port: http
+startupProbe:
+  httpGet:
+    path: /health
+    port: http
+  initialDelaySeconds: 30 # Adjust this value according to the startup time of the application
+  periodSeconds: 10 # Adjust this value according to the startup time of the application
+  failureThreshold: 20 # Adjust this value according to the startup time of the application

charts/open-webui/values.yaml

@@ -106,6 +106,39 @@ redis-cluster:
     # -- Number of Redis replica instances
     replicaCount: 3
 
+rag:
+  # -- Enable RAG
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#retrieval-augmented-generation-rag
+  enabled: false
+  # -- Vector database configuration
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#vector_db
+  vectorDB: ""
+  # -- Embedding engine to use for RAG with env `RAG_EMBEDDING_ENGINE`: ""(empty), "ollama", "openai"
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#rag_embedding_engine
+  embeddingEngine: ""
+  # -- Embedding model to use for RAG with env `RAG_EMBEDDING_MODEL`
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#rag_embedding_model
+  embeddingModel: ""
+
+milvus:
+  # -- Enable Milvus installation. Deploys a Milvus cluster/standalone with subchart 'milvus' from zilliztech
+  # ref: https://github.com/zilliztech/milvus-helm/tree/master/charts/milvus
+  enabled: false
+  # -- Milvus fullname override (recommended to be 'open-webui-milvus')
+  # - In this case, the Milvus uri will be 'http://[username:password@]open-webui-milvus:19530'
+  fullnameOverride: open-webui-milvus
+  # -- Active Milvus URI for RAG with env `MILVUS_URI`. If there is credentials in the uri, it will be used to connect to the Milvus server.
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_uri
+  uri: "http://open-webui-milvus:19530"
+  # -- Example `milvus.uri` with credentials (Not recommended for production. Use `env` with `secretKeyRef` instead)
+  # uri: "http://username:password@open-webui-milvus:19530"
+  # -- Active Milvus database for RAG with env `MILVUS_DB`
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_db
+  db: default
+  # -- Active Milvus token for RAG with env `MILVUS_TOKEN`
+  # ref: https://docs.openwebui.com/getting-started/env-configuration#milvus_token
+  token: {}
+
 # -- Value of cluster domain
 clusterDomain: cluster.local