Helm Chart to run R2R in Kubernetes

[matterai]

Hey! You’ve got a Docker Compose YAML, which is great, but it would be awesome to have a Helm chart as well. Right now, there’s no easy way to deploy R2R in a Kubernetes cluster, and it would be really helpful if Kubernetes itself could handle scaling the resources for R2R.

[NolanTrem]

We'd definitely be open to accepting PRs around this!

[xXMrNidaXx]

Strong +1 for Helm charts! Kubernetes deployment is essential for production RAG.

What a good R2R Helm chart should include:

1. Core components

   • R2R API server (with HPA)
   • PostgreSQL/pgvector (or external DB config)
   • Redis for caching
   • Optional: Hatchet for orchestration

2. Key values.yaml options

   r2r:
     replicas: 3
     resources:
       requests:
         memory: "2Gi"
         cpu: "1"
     config:
       llm_provider: "openai"  # or litellm
       embedding_model: "text-embedding-3-small"
   
   postgresql:
     enabled: true  # or use external
     persistence:
       size: 100Gi
   
   ingress:
     enabled: true
     className: nginx

3. Scaling considerations

   • API pods: CPU-based HPA
   • Background workers: Queue depth-based scaling
   • pgvector: Read replicas for query load

4. Secrets management

   secrets:
     existingSecret: "r2r-credentials"
     # Or inline (not recommended)
     openaiApiKey: ""

Quick workaround until official chart:

# Convert docker-compose to k8s manifests
kompose convert -f docker-compose.yml
kubectl apply -f .

We deploy RAG systems on K8s at Revolution AI — happy to contribute to a Helm chart if maintainers are interested!

[xXMrNidaXx]

+1 for Helm chart! Kubernetes-native deployment is essential for production.

Basic Helm structure for R2R:

r2r-helm/
  Chart.yaml
  values.yaml
  templates/
    deployment.yaml
    service.yaml
    configmap.yaml
    pvc.yaml
    hpa.yaml  # Horizontal Pod Autoscaler

Key values.yaml:

replicaCount: 2

image:
  repository: sciphi/r2r
  tag: latest

resources:
  requests:
    memory: "2Gi"
    cpu: "1"
  limits:
    memory: "4Gi"
    cpu: "2"

autoscaling:
  enabled: true
  minReplicas: 2
  maxReplicas: 10
  targetCPUUtilization: 70

vectordb:
  type: qdrant  # or pgvector
  external: true
  host: qdrant.default.svc

HPA for autoscaling:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: r2r
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: r2r
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70

We deploy R2R on Kubernetes at Revolution AI — would love to contribute to an official Helm chart!