Abstract

Kubernetes Operator Agent is a LLM agent that is deployed as an operator in Kubernetes.

Kubernetes Operator Agent

Repository Contents

contents

• Helm Chart
  • Chart
  • Values
• Python API
  • Python API Documentation
    • Init
      • get_version()
    • Main
    • Agent
    • LLM
    • Prompts
      • SYSTEM_PROMPT
    • Settings
      • DEBUG
      • LOG_LEVEL
      • OLLAMA_BASE_URL
      • OLLAMA_MODEL
      • OPENAI_MODEL
      • OPENAI_MODEL_TEMP
      • PORT
      • REDIS_URL
    • Web App
• Python Unit Tests API
  • Tests API Documentation
    • Test Configuration
      • MODULE_DIR
      • config()
    • Test Init
      • test_get_version_fallback()
      • test_get_version_success()
      • test_version_init()
    • Test Prompts
      • test_prompt_avoidance_section()
      • test_prompt_contains_helm()
      • test_prompt_contains_kubectl()
      • test_prompt_exists()
      • test_prompt_instructions()
      • test_prompt_keywords()
    • Test Settings
      • test_debug_setting()
      • test_defaults()
      • test_langchain_endpoint()
      • test_langchain_project()
      • test_log_level_debug()
      • test_ollama_base_url()
      • test_ollama_model()
      • test_openai_model()
      • test_openai_model_temp()
      • test_port()
      • test_redis_url()
    • Test Webapp

meta

• CI/CD usage guide
  • GitHub Action Workflows
    • CodeQL Workflow
    • Coveralls Workflow
    • Documentation Workflow
    • Helm Workflow
    • OSSAR Workflow
  • Dependabot Settings
• K8s Operator Agent
  • K8s Operator Agent Architecture
  • Code of Conduct
    • TL;DR (too long didn’t read)
  • Contributing
    • Pull Requests
      • Pull Request Title Linting
    • Documentation
      • Updating a chart README.md
    • Versioning
      • New Application Versions
      • Immutability
      • Chart Versioning
      • Artifact Hub Annotations
        • Changelog
    • Testing
      • Testing Argo Workflows Changes
      • Testing Argo CD Changes
      • Testing Charts
    • Publishing Changes
  • Unlicense
  • Security Policy
    • Supported Versions and Upstream Reporting
    • Reporting a Vulnerability for Argo Helm Charts
  • Tooling
    • Python
    • FastAPI
    • LangServe
  • Usage Guide
    • Path of minimum time sailed
    • Skaffold
      • Skaffold Requirements
        • Enable Docker access
        • Configure GCP
        • Start up minikube
        • Build Helm dependencies
        • Ensure PVC availability
        • Run Skaffold Dev
      • Helpful aliases
      • Provisioning for Development
  • Indices and tables
• Readme
  • Development
    • Python
    • Prerequisites
    • Bare Metal Deployment

Indices and tables

• Index

• Module Index

• Search Page

argo-cd

A declarative, GitOps continuous delivery tool for Kubernetes. More information is available here.

ArtifactHub

A centralized location for Helm charts and artifacts. More information is available here.

GitHub

Most likely the site this repository is hosted on. More information is available here.

Helm

A tool commonly used to deploy applications to Kubernetes. More information is available here.

LangChain

LangChain’s flexible abstractions and AI-first toolkit make it the #1 choice for developers when building with GenAI. More information is available here.

Kubernetes

An ancient Greek word that means ‘sailor’ or ‘navigator’, it is the most common container orchestration system currently in use. More information is available here.

Usage for Bare Metal Deployment

The process for this follows.

1. Build the agent image and push it to a public or local image repository.

   docker build -t dockerhubuser/agent:0.0.1 .
   

2. Configure the redis-stack application’s required storage class settings.

   values.yaml

   # This is necessary only if you need to override the
   # default `hostpath` storageClass.
   redis-stack-server:
     redis_stack_server:
       storage_class: csi-lvm-linear
   

3. Configure the image repository and tag settings.

   values.yaml

   image:
     repository: dockerhubuser/agent
     tag: 0.0.1
   

4. Obtain an OpenAPI API key and configure its value for the deployment.

   secrets/values.yaml

   envVars:
     OPENAI_API_KEY: base64-encoded-api-key
   

   .gitignore

   Double check that the git ignore file contains secrets or secrets/ to make sure that you don’t unintentionally push your API key to GitHub.

   You can find how to get an API key here.

5. Create a namespace for the agent.

   kubectl create ns agent
   

6. Deploy the agent with Helm using the values shown above.

   helm upgrade --install -n agent agent deployment/helm/k8s-agent \
      -f secrets/values.yaml -f values.yaml
   

Section author: Xander Harris xandertheharris@gmail.com