Update charts for tantivy search container

windmill-labs · Oct 16, 2024 · 813cf32 · 813cf32
1 parent 5dd05bb
commit 813cf32
Show file tree

Hide file tree

Showing 3 changed files with 255 additions and 0 deletions.
diff --git a/charts/windmill/templates/indexer.yaml b/charts/windmill/templates/indexer.yaml
@@ -0,0 +1,193 @@
+{{- if and .Values.enterprise.enabled .Values.windmill.indexer.enabled -}}
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: windmill-indexer
+  labels:
+    app: windmill-indexer
+    app.kubernetes.io/name: windmill-indexer
+    chart: {{ template "windmill.chart" . }}
+    release: {{ .Release.Name }}
+    heritage: {{ .Release.Service }}
+spec:
+  replicas: 1
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 3
+      maxUnavailable: 0
+  selector:
+    matchLabels:
+      app: windmill-indexer
+  template:
+    metadata:
+      labels:
+        app: windmill-indexer
+        app.kubernetes.io/name: windmill-indexer
+        container: windmill-indexer
+{{- with .Values.windmill.indexer.labels }}
+{{ toYaml . | indent 8 }}
+{{- end }}
+{{- with .Values.windmill.indexer.annotations }}
+      annotations:
+{{ toYaml . | indent 8 }}
+  {{- end }}
+    spec:
+      {{ if .Values.windmill.imagePullSecrets }}
+      imagePullSecrets:
+      - name: {{ .Values.windmill.imagePullSecrets }}
+      {{ end }}
+      containers:
+      - name: windmill-indexer
+        securityContext:
+      {{- with .Values.windmill.indexer.containerSecurityContext }}
+      {{- toYaml . | nindent 10 }}
+      {{- end }}
+        image: {{ default "ghcr.io/windmill-labs/windmill-ee" .Values.windmill.image }}:{{ default "latest" .Values.windmill.tag }}
+        imagePullPolicy: Always
+        ports:
+        - containerPort: 8000
+        readinessProbe:
+          timeoutSeconds: 1
+          successThreshold: 1
+          failureThreshold: 1
+          httpGet:
+            host:
+            scheme: HTTP
+            path: /
+            httpHeaders:
+            - name: Host
+              value: localhost
+            port: 8000
+          initialDelaySeconds: 5
+          periodSeconds: 5
+        env:
+        {{- with .Values.windmill.indexer.extraEnv }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{ if .Values.enterprise.enabled }}
+        - name : "METRICS_ADDR"
+          value: "true"
+        {{ end }}
+        {{ if .Values.windmill.databaseUrlSecretName }}
+        - name: "DATABASE_URL"
+          valueFrom:
+            secretKeyRef:
+              name: "{{ .Values.windmill.databaseUrlSecretName }}"
+              key: "{{ .Values.windmill.databaseUrlSecretKey }}"
+        {{ else }}
+        - name: "DATABASE_URL"
+          value: "{{ .Values.windmill.databaseUrl }}"
+        {{ end }}
+        - name: "BASE_URL"
+          value: "{{ .Values.windmill.baseProtocol }}://{{ .Values.windmill.baseDomain }}"
+        - name: "RUST_LOG"
+          value: "{{ .Values.windmill.rustLog }}"
+        - name: "MODE"
+          value: "indexer"
+        - name: "JSON_FMT"
+          value: "true"
+        {{ if .Values.windmill.instanceEventsWebhook }}
+        - name: "INSTANCE_EVENTS_WEBHOOK"
+          value: "true"
+        {{ end }}
+        {{ if .Values.enterprise.enabled }}
+        {{ if .Values.enterprise.licenseKeySecretName }}
+        - name: "LICENSE_KEY"
+          valueFrom:
+            secretKeyRef:
+              name: {{ .Values.enterprise.licenseKeySecretName }}
+              key: {{ .Values.enterprise.licenseKeySecretKey }}
+        {{ else }}
+        - name: "LICENSE_KEY"
+          value: "{{ .Values.enterprise.licenseKey }}"
+        {{ end }}
+        {{ if .Values.enterprise.samlMetadata }}
+        - name: "SAML_METADATA"
+          value: "{{ .Values.enterprise.samlMetadata }}"
+        {{ end }}
+        {{ if $.Values.enterprise.enabledS3DistributedCache }}
+        - name: "S3_CACHE_BUCKET"
+          value: "{{ $.Values.enterprise.s3CacheBucket }}"
+        {{ end }}
+        {{if .Values.enterprise.scimTokenSecretName}}
+        - name: "SCIM_TOKEN"
+          valueFrom:
+            secretKeyRef:
+              name: "{{ .Values.enterprise.scimTokenSecretName }}"
+              key: "{{ .Values.enterprise.scimTokenSecretKey }}"
+        {{ else if .Values.enterprise.scimToken}}
+        - name: "SCIM_TOKEN"
+          value: "{{ .Values.enterprise.scimToken }}"
+        {{ end }}
+        {{ end }}
+        {{ if .Values.windmill.openaiAzureBasePath}}
+        - name: OPENAI_AZURE_BASE_PATH
+          value: "{{ .Values.windmill.openaiAzureBasePath }}"
+        {{ end }}
+        {{ if .Values.windmill.cookieDomain }}
+        - name: "COOKIE_DOMAIN"
+          value: "{{ .Values.windmill.cookieDomain }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyMaxIndexedJobLogSize__mb }}
+        - name: "TANTIVY_MAX_INDEXED_JOB_LOG_SIZE__MB"
+          value: "{{ .Values.windmill.indexer.tantivyMaxIndexedJobLogSize__mb }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyS3BackupPeriod__s }}
+        - name: "TANTIVY_S3_BACKUP_PERIOD__S"
+          value: "{{ .Values.windmill.indexer.tantivyS3BackupPeriod__s }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyIndexWriterMemoryBudget__mb }}
+        - name: "TANTIVY_INDEX_WRITER_MEMORY_BUDGET__MB"
+          value: "{{ .Values.windmill.indexer.tantivyIndexWriterMemoryBudget__mb }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyRefreshIndexPeriod__s }}
+        - name: "TANTIVY_REFRESH_INDEX_PERIOD__S"
+          value: "{{ .Values.windmill.indexer.tantivyRefreshIndexPeriod__s }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyDocCommitMaxBatchSize }}
+        - name: "TANTIVY_DOC_COMMIT_MAX_BATCH_SIZE"
+          value: "{{ .Values.windmill.indexer.tantivyDocCommitMaxBatchSize }}"
+        {{ end }}
+        {{ if .Values.windmill.indexer.tantivyShowMemoryEvery }}
+        - name: "TANTIVY_SHOW_MEMORY_EVERY"
+          value: "{{ .Values.windmill.indexer.tantivyShowMemoryEvery }}"
+        {{ end }}
+
+        resources:
+{{ toYaml .Values.windmill.indexer.resources | indent 12 }}
+    {{- with .Values.windmill.indexer.nodeSelector }}
+      nodeSelector:
+{{ toYaml . | indent 8 }}
+    {{- end }}
+    {{- with .Values.windmill.indexer.affinity }}
+      affinity:
+{{ toYaml . | indent 8 }}
+    {{- end }}
+    {{- with .Values.windmill.indexer.tolerations }}
+      tolerations:
+{{ toYaml . | indent 8 }}
+    {{- end }}
+    {{- with .Values.windmill.indexer.podSecurityContext }}
+      securityContext:
+{{ toYaml . | indent 8 }}
+    {{- end }}
+    {{- with .Values.windmill.indexer.securityContext }}
+      securityContext:
+{{ toYaml . | indent 8 }}
+    {{- end }}
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: windmill-indexer
+spec:
+  ports:
+    - port: 8000
+      protocol: TCP
+      targetPort: 8000
+  selector:
+    app.kubernetes.io/name: windmill-indexer
+  sessionAffinity: ClientIP
+  type: ClusterIP
+{{- end -}}
diff --git a/charts/windmill/templates/ingress.yaml b/charts/windmill/templates/ingress.yaml
@@ -43,6 +43,18 @@ spec:
                 name: windmill-multiplayer
                 port:
                   number: 3002
+    {{ if .Values.windmill.indexer.enabled }}
+    - host: {{ .Values.windmill.baseDomain | quote }}
+      http:
+        paths:
+          - path: /api/srch/
+            pathType: Prefix
+            backend:
+              service:
+                name: windmill-indexer
+                port:
+                  number: 8000
+    {{ end }}
     {{ if .Values.hub.enabled }}
     - host: {{ .Values.hub.baseDomain | quote }}
       http:

diff --git a/charts/windmill/values.yaml b/charts/windmill/values.yaml
@@ -369,6 +369,56 @@ windmill:
       # -- target CPU utilization
       targetCPUUtilizationPercentage: 80
 
+  # indexer configuration
+  indexer:
+    tag: "latest"
+
+    enabled: true
+
+    # -- Annotations to apply to the pods
+    annotations: {}
+
+    # -- Annotations to apply to the pods
+    labels: {}
+
+    # -- Node selector to use for scheduling the pods
+    nodeSelector: {}
+
+    # -- Tolerations to apply to the pods
+    tolerations: []
+
+    # -- legacy, use podSecurityContext instead
+    securityContext:
+    # -- Security context to apply to the pods
+    podSecurityContext:
+      # -- run as user. The default is 0 for root user
+      runAsUser: 0
+      # -- run explicitly as a non-root user. The default is false.
+      runAsNonRoot: false
+    containerSecurityContext: {}
+
+    # -- Affinity rules to apply to the pods
+    affinity: {}
+
+    # -- Resource limits and requests for the pods
+    resources: {}
+
+    # -- job logs bigger than this will be truncated before indexing
+    tantivyMaxIndexedJobLogSize__mb: 1
+    # -- how often to backup the index into object storage in seconds
+    tantivyS3BackupPeriod__s: 3600
+    # -- higher budget for higher indexing throughput
+    tantivyIndexWriterMemoryBudget__mb: 500
+    # -- how often to start indexing new jobs
+    tantivyRefreshIndexPeriod__s: 300
+    # -- how many documents to batch in one commit
+    tantivyDocCommitMaxBatchSize: 100000
+    # -- log memory usage and progress every so many documents indexed
+    tantivyShowMemoryEvery: 10000
+
+    # -- Extra environment variables to apply to the pods
+    extraEnv: []
+
 ingress:
   # -- enable/disable included ingress resource
   enabled: true