Helm Charts Reference

Helm chart reference for LLM Platform Kubernetes deployment

Overview

The LLM Platform uses Helm charts for Kubernetes deployment, providing templated, version-controlled infrastructure as code.

Chart Version: 1.0.0 Kubernetes: 1.28+ Helm: 3.12+

Chart Structure

infrastructure/helm-chart/
 Chart.yaml               # Chart metadata
 values.yaml             # Default values
 values-dev.yaml         # Development overrides
 values-staging.yaml     # Staging overrides
 values-prod.yaml        # Production overrides
 README.md
 templates/
     _helpers.tpl        # Template helpers
     NOTES.txt           # Post-install notes
     deployments/
        drupal.yaml
        llm-gateway.yaml
        agent-mesh.yaml
        agent-tracer.yaml
        agent-buildkit.yaml
     services/
        drupal.yaml
        llm-gateway.yaml
        agent-mesh.yaml
        postgres.yaml
     statefulsets/
        postgres.yaml
        redis.yaml
        qdrant.yaml
        neo4j.yaml
        mongodb.yaml
     configmaps/
        drupal-config.yaml
        llm-gateway-config.yaml
        nginx-config.yaml
     secrets/
        api-keys.yaml
     ingress.yaml
     hpa.yaml
     pvc.yaml
     serviceaccount.yaml

Chart.yaml

apiVersion: v2
name: llm-platform
description: Enterprise AI Orchestration Platform
type: application
version: 1.0.0
appVersion: "1.0.0"
keywords:
  - ai
  - llm
  - drupal
  - kubernetes
home: https://gitlab.com/blueflyio/agent-platform/demos/llm-platform-demo
sources:
  - https://gitlab.com/blueflyio/agent-platform/demos/llm-platform-demo
maintainers:
  - name: Bluefly Team
    email: dev@bluefly.io
dependencies:
  - name: postgresql
    version: 12.x.x
    repository: https://charts.bitnami.com/bitnami
    condition: postgres.enabled
  - name: redis
    version: 18.x.x
    repository: https://charts.bitnami.com/bitnami
    condition: redis.enabled

Values Files

values.yaml (Defaults)

# Global settings
global:
  environment: development
  domain: llm-platform.local

# Drupal Platform
drupal:
  enabled: true
  replicaCount: 1
  image:
    repository: registry.gitlab.com/bluefly/llm-platform
    pullPolicy: IfNotPresent
    tag: "latest"

  service:
    type: ClusterIP
    port: 80

  resources:
    requests:
      cpu: 500m
      memory: 1Gi
    limits:
      cpu: 1000m
      memory: 2Gi

  autoscaling:
    enabled: false
    minReplicas: 1
    maxReplicas: 10
    targetCPUUtilizationPercentage: 70

# LLM Gateway
llmGateway:
  enabled: true
  replicaCount: 1
  image:
    repository: registry.gitlab.com/bluefly/llm-gateway
    tag: "1.0.0"

  service:
    type: ClusterIP
    port: 4000

  env:
    ANTHROPIC_API_KEY: ""
    OPENAI_API_KEY: ""
    REDIS_URL: "redis://redis:6379"

# Agent Mesh
agentMesh:
  enabled: true
  replicaCount: 1
  image:
    repository: registry.gitlab.com/bluefly/agent-mesh
    tag: "1.0.0"

  service:
    type: ClusterIP
    httpPort: 3005
    grpcPort: 50051

# Agent Tracer
agentTracer:
  enabled: true
  replicaCount: 1
  image:
    repository: registry.gitlab.com/bluefly/agent-tracer
    tag: "1.0.0"

  service:
    type: ClusterIP
    port: 3007

# PostgreSQL
postgres:
  enabled: true
  auth:
    username: llm_user
    database: llm_platform
    password: ""
  primary:
    persistence:
      enabled: true
      size: 10Gi

# Redis
redis:
  enabled: true
  architecture: standalone
  auth:
    enabled: true
    password: ""
  master:
    persistence:
      enabled: true
      size: 1Gi

# Qdrant
qdrant:
  enabled: true
  replicaCount: 1
  image:
    repository: qdrant/qdrant
    tag: "latest"
  persistence:
    enabled: true
    size: 10Gi

# Ingress
ingress:
  enabled: false
  className: nginx
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
  hosts:
    - host: llm-platform.local
      paths:
        - path: /
          pathType: Prefix
  tls: []

values-dev.yaml

global:
  environment: development
  domain: llm-platform.dev.local

drupal:
  replicaCount: 1
  autoscaling:
    enabled: false

  resources:
    requests:
      cpu: 250m
      memory: 512Mi
    limits:
      cpu: 500m
      memory: 1Gi

llmGateway:
  replicaCount: 1

postgres:
  primary:
    persistence:
      size: 5Gi

qdrant:
  persistence:
    size: 5Gi

ingress:
  enabled: true
  hosts:
    - host: llm-platform.dev.local

values-staging.yaml

global:
  environment: staging
  domain: staging.llm-platform.example.com

drupal:
  replicaCount: 2
  autoscaling:
    enabled: true
    minReplicas: 2
    maxReplicas: 5

  resources:
    requests:
      cpu: 500m
      memory: 1Gi
    limits:
      cpu: 1000m
      memory: 2Gi

llmGateway:
  replicaCount: 2

postgres:
  primary:
    persistence:
      size: 50Gi

qdrant:
  persistence:
    size: 50Gi

ingress:
  enabled: true
  hosts:
    - host: staging.llm-platform.example.com
  tls:
    - secretName: staging-tls
      hosts:
        - staging.llm-platform.example.com

values-prod.yaml

global:
  environment: production
  domain: llm-platform.example.com

drupal:
  replicaCount: 3
  autoscaling:
    enabled: true
    minReplicas: 3
    maxReplicas: 10
    targetCPUUtilizationPercentage: 70

  resources:
    requests:
      cpu: 1000m
      memory: 2Gi
    limits:
      cpu: 2000m
      memory: 4Gi

llmGateway:
  replicaCount: 3
  autoscaling:
    enabled: true
    minReplicas: 3
    maxReplicas: 10

agentMesh:
  replicaCount: 2

postgres:
  primary:
    persistence:
      enabled: true
      size: 100Gi
      storageClass: gp3
    resources:
      requests:
        cpu: 2000m
        memory: 4Gi
      limits:
        cpu: 4000m
        memory: 8Gi

redis:
  architecture: replication
  replica:
    replicaCount: 2
  master:
    persistence:
      size: 10Gi

qdrant:
  replicaCount: 3
  persistence:
    enabled: true
    size: 100Gi

ingress:
  enabled: true
  annotations:
    nginx.ingress.kubernetes.io/ssl-redirect: "true"
    nginx.ingress.kubernetes.io/force-ssl-redirect: "true"
  hosts:
    - host: llm-platform.example.com
      paths:
        - path: /
          pathType: Prefix
  tls:
    - secretName: llm-platform-tls
      hosts:
        - llm-platform.example.com

Template Examples

Deployment Template

# templates/deployments/drupal.yaml
{{- if .Values.drupal.enabled }}
apiVersion: apps/v1
kind: Deployment
metadata:
  name: {{ include "llm-platform.fullname" . }}-drupal
  labels:
    {{- include "llm-platform.labels" . | nindent 4 }}
    app.kubernetes.io/component: drupal
spec:
  {{- if not .Values.drupal.autoscaling.enabled }}
  replicas: {{ .Values.drupal.replicaCount }}
  {{- end }}
  selector:
    matchLabels:
      {{- include "llm-platform.selectorLabels" . | nindent 6 }}
      app.kubernetes.io/component: drupal
  template:
    metadata:
      labels:
        {{- include "llm-platform.selectorLabels" . | nindent 8 }}
        app.kubernetes.io/component: drupal
    spec:
      containers:
      - name: drupal
        image: "{{ .Values.drupal.image.repository }}:{{ .Values.drupal.image.tag }}"
        imagePullPolicy: {{ .Values.drupal.image.pullPolicy }}
        ports:
        - name: http
          containerPort: 80
          protocol: TCP
        env:
        - name: DRUPAL_DATABASE_HOST
          value: {{ include "llm-platform.fullname" . }}-postgres
        - name: DRUPAL_DATABASE_PORT
          value: "5432"
        - name: DRUPAL_DATABASE_NAME
          value: {{ .Values.postgres.auth.database }}
        - name: DRUPAL_DATABASE_USERNAME
          value: {{ .Values.postgres.auth.username }}
        - name: DRUPAL_DATABASE_PASSWORD
          valueFrom:
            secretKeyRef:
              name: {{ include "llm-platform.fullname" . }}-postgres
              key: password
        livenessProbe:
          httpGet:
            path: /
            port: http
          initialDelaySeconds: 30
          periodSeconds: 10
        readinessProbe:
          httpGet:
            path: /
            port: http
          initialDelaySeconds: 10
          periodSeconds: 5
        resources:
          {{- toYaml .Values.drupal.resources | nindent 12 }}
{{- end }}

Service Template

# templates/services/drupal.yaml
{{- if .Values.drupal.enabled }}
apiVersion: v1
kind: Service
metadata:
  name: {{ include "llm-platform.fullname" . }}-drupal
  labels:
    {{- include "llm-platform.labels" . | nindent 4 }}
spec:
  type: {{ .Values.drupal.service.type }}
  ports:
  - port: {{ .Values.drupal.service.port }}
    targetPort: http
    protocol: TCP
    name: http
  selector:
    {{- include "llm-platform.selectorLabels" . | nindent 4 }}
    app.kubernetes.io/component: drupal
{{- end }}

HPA Template

# templates/hpa.yaml
{{- if .Values.drupal.autoscaling.enabled }}
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: {{ include "llm-platform.fullname" . }}-drupal
  labels:
    {{- include "llm-platform.labels" . | nindent 4 }}
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: {{ include "llm-platform.fullname" . }}-drupal
  minReplicas: {{ .Values.drupal.autoscaling.minReplicas }}
  maxReplicas: {{ .Values.drupal.autoscaling.maxReplicas }}
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: {{ .Values.drupal.autoscaling.targetCPUUtilizationPercentage }}
{{- end }}

Helm Commands

Install

# Development
helm install llm-platform ./infrastructure/helm-chart \
  --namespace llm-platform \
  --create-namespace \
  --values infrastructure/helm-chart/values-dev.yaml

# Staging
helm install llm-platform ./infrastructure/helm-chart \
  --namespace llm-platform \
  --create-namespace \
  --values infrastructure/helm-chart/values-staging.yaml

# Production
helm install llm-platform ./infrastructure/helm-chart \
  --namespace llm-platform \
  --create-namespace \
  --values infrastructure/helm-chart/values-prod.yaml

Upgrade

# Upgrade with new values
helm upgrade llm-platform ./infrastructure/helm-chart \
  --namespace llm-platform \
  --values infrastructure/helm-chart/values-prod.yaml

# Upgrade with set values
helm upgrade llm-platform ./infrastructure/helm-chart \
  --namespace llm-platform \
  --set drupal.replicaCount=5

Rollback

# List releases
helm history llm-platform -n llm-platform

# Rollback to previous version
helm rollback llm-platform -n llm-platform

# Rollback to specific revision
helm rollback llm-platform 2 -n llm-platform

Uninstall

# Uninstall release
helm uninstall llm-platform -n llm-platform

# Keep history
helm uninstall llm-platform -n llm-platform --keep-history

Testing Charts

# Lint chart
helm lint ./infrastructure/helm-chart

# Dry run
helm install llm-platform ./infrastructure/helm-chart \
  --dry-run \
  --debug \
  --namespace llm-platform

# Template rendering
helm template llm-platform ./infrastructure/helm-chart \
  --values infrastructure/helm-chart/values-prod.yaml

Related Documentation

• Cluster Setup
• Golden Workflow CI/CD
• System Overview