Skip to content

Install Python SDK

Embed HolmesGPT in your own applications with the Python API for programmatic root cause analysis.

Installation

pip install holmesgpt

pipenv install holmesgpt

poetry add holmesgpt

Quick Start

Basic Usage

from holmes import ask_holmes

# Simple question
result = ask_holmes("what pods are failing in production?")
print(result)

# With specific context
result = ask_holmes(
    "analyze this pod failure",
    context_files=["pod.yaml", "logs.txt"]
)
print(result)

Advanced Configuration

from holmes.core import HolmesGPT
from holmes.config import Config

# Configure with custom settings
config = Config(
    ai_provider="openai",
    api_key="your-api-key",
    model="gpt-4",
    max_tokens=2000
)

# Initialize Holmes
holmes = HolmesGPT(config)

# Run investigation
result = holmes.investigate("what's wrong with my deployment?")

from holmes.core import HolmesGPT
from holmes.toolsets import CustomToolset

# Define custom toolset
class MyCustomToolset(CustomToolset):
    def get_custom_data(self):
        # Your custom data collection logic
        return {"custom_metrics": "data"}

# Configure Holmes with custom toolset
holmes = HolmesGPT()
holmes.add_toolset(MyCustomToolset())

result = holmes.investigate("analyze using my custom data")

API Reference

Core Functions

ask_holmes(question, **kwargs)

Ask Holmes a question synchronously.

Parameters: - question (str): The question to ask - context_files (list, optional): List of file paths for context - toolsets (list, optional): Custom toolsets to use - config (Config, optional): Custom configuration

Returns: - str: Holmes' analysis and recommendations

ask_holmes_async(question, **kwargs)

Async version of ask_holmes().

HolmesGPT(config=None)

Main Holmes class for advanced usage.

Methods: - investigate(question): Run investigation - add_toolset(toolset): Add custom toolset - set_context(context): Set investigation context

Configuration Options

from holmes.config import Config

config = Config(
    # AI Provider settings
    ai_provider="openai",  # "openai", "anthropic", "bedrock", "vertex"
    api_key="your-key",
    model="gpt-4",

    # Investigation settings
    max_tokens=2000,
    temperature=0.1,
    enable_context=True,

    # Kubernetes settings
    kubeconfig_path="~/.kube/config",
    default_namespace="default",

    # Custom toolsets
    custom_toolsets=[],
    enabled_toolsets=["kubernetes", "prometheus"]
)

Environment Variables

Set these environment variables for configuration:

# AI Provider (choose one)
export OPENAI_API_KEY="your-openai-key"
export ANTHROPIC_API_KEY="your-anthropic-key"
export GOOGLE_API_KEY="your-google-key"

# Optional: Custom configuration
export HOLMES_CONFIG_PATH="/path/to/config.yaml"
export HOLMES_LOG_LEVEL="INFO"

Examples Repository

Check out our examples repository for:

  • Flask integration - Web app with Holmes
  • Slack bot - Custom Slack integration
  • Jupyter notebooks - Data analysis workflows
  • CI/CD pipelines - GitHub Actions examples

Next Steps

Need Help?