Skip to content

Install Helm Chart

Deploy HolmesGPT as a service in your Kubernetes cluster with an HTTP API.

When to use the Helm chart?

Most users should use the CLI or UI/TUI instead. Using the Helm chart is only recommended if you're building a custom integration over an HTTP API.

Prerequisites

  • Kubernetes cluster
  • Helm
  • kubectl configured to access your cluster
  • Supported AI Provider API key.

RBAC Permissions

The Helm chart automatically creates a ServiceAccount with ClusterRole permissions required for HolmesGPT to analyze your cluster. For details on required permissions, see Kubernetes Permissions.

Installation

  1. Add the Helm repository: 

    helm repo add robusta https://robusta-charts.storage.googleapis.com
helm repo update

  2. Create values.yaml file:

    Create a values.yaml file to configure HolmesGPT with your models using the modelList approach:

    # values.yaml
additionalEnvVars:
- name: OPENAI_API_KEY
  value: "your-openai-api-key"
# Or load from secret:
# - name: OPENAI_API_KEY
#   valueFrom:
#     secretKeyRef:
#       name: holmes-secrets
#       key: openai-api-key

modelList:
  gpt-4.1:
    api_key: "{{ env.OPENAI_API_KEY }}"
    model: openai/gpt-4.1
    temperature: 0
  gpt-5:
    api_key: "{{ env.OPENAI_API_KEY }}"
    model: openai/gpt-5

    # values.yaml
additionalEnvVars:
- name: ANTHROPIC_API_KEY
  value: "your-anthropic-api-key"
# Or load from secret:
# - name: ANTHROPIC_API_KEY
#   valueFrom:
#     secretKeyRef:
#       name: holmes-secrets
#       key: anthropic-api-key

modelList:
  claude-sonnet:
    api_key: "{{ env.ANTHROPIC_API_KEY }}"
    model: anthropic/claude-3-5-sonnet-20241022
    temperature: 0

    # values.yaml
additionalEnvVars:
- name: AZURE_API_KEY
  value: "your-azure-api-key"
- name: AZURE_API_BASE
  value: "https://your-resource.openai.azure.com/"
- name: AZURE_API_VERSION
  value: "2024-02-15-preview"
# Or load from secret:
# - name: AZURE_API_KEY
#   valueFrom:
#     secretKeyRef:
#       name: holmes-secrets
#       key: azure-api-key
# - name: AZURE_API_BASE
#   valueFrom:
#     secretKeyRef:
#       name: holmes-secrets
#       key: azure-api-base

modelList:
  azure-gpt4:
    api_key: "{{ env.AZURE_API_KEY }}"
    model: azure/your-deployment-name
    api_base: "{{ env.AZURE_API_BASE }}"
    api_version: "{{ env.AZURE_API_VERSION }}"
    temperature: 0

    # values.yaml
additionalEnvVars:
- name: OPENAI_API_KEY
  value: "your-openai-api-key"
- name: ANTHROPIC_API_KEY
  value: "your-anthropic-api-key"
# Or load from secrets (recommended)

modelList:
  gpt-4.1:
    api_key: "{{ env.OPENAI_API_KEY }}"
    model: openai/gpt-4.1
    temperature: 0
  claude-sonnet:
    api_key: "{{ env.ANTHROPIC_API_KEY }}"
    model: anthropic/claude-3-5-sonnet-20241022
    temperature: 0
  gpt-5:
    api_key: "{{ env.OPENAI_API_KEY }}"
    model: openai/gpt-5

    Configuration Guide: Each AI provider requires different environment variables. See the AI Providers documentation for the specific environment variables needed for your chosen provider, then add them to the additionalEnvVars section as shown above. For a complete list of all environment variables, see the Environment Variables Reference. For advanced multiple provider setup, see Using Multiple Providers.

  3. Install HolmesGPT: 

    helm install holmesgpt robusta/holmes -f values.yaml

Usage

After installation, test the service with a simple API call:

# Port forward to access the service locally
# Note: Service name is {release-name}-holmes
kubectl port-forward svc/holmesgpt-holmes 8080:80

# If you used a different release name or namespace:
# kubectl port-forward svc/{your-release-name}-holmes 8080:80 -n {your-namespace}

# Test with a basic question using modelList model name
curl -X POST http://localhost:8080/api/chat \
  -H "Content-Type: application/json" \
  -d '{"ask": "list pods in namespace default?", "model": "gpt-4o-mini"}'

# Using a different model from your modelList
curl -X POST http://localhost:8080/api/chat \
  -H "Content-Type: application/json" \
  -d '{"ask": "list pods in namespace default?", "model": "claude-sonnet"}'

Note: Responses may take some time when HolmesGPT needs to gather large amounts of data to answer your question. Streaming APIs are coming soon to stream results.

For complete API documentation, see the HTTP API Reference.

Upgrading

helm repo update
helm upgrade holmesgpt robusta/holmes -f values.yaml

Uninstalling

helm uninstall holmesgpt

Need Help?