Manage AI Agents

How to create and manage AI agents for cluster intelligence and automation.
Tags:
Categories:

Faros integrates AI agents powered by LLMs to provide intelligent cluster analysis, recommendations, and automation capabilities. AI agents can be deployed to analyze your clusters and provide actionable insights.

Prerequisites

  • Ensure you have installed the Faros CLI. If not, refer to the CLI installation guide.
  • Authenticate with Faros: kubectl faros login
  • An API key for your chosen AI backend (e.g., OpenAI)

Listing AI Agents

View all AI agents in your namespace:

kubectl faros ai-agents list
# or simply
kubectl faros ai-agents

This displays:

  • NAME: Agent name
  • BACKEND: AI backend (e.g., openai)
  • MODEL: Model being used (e.g., gpt-4)
  • PHASE: Current status (Pending, Initializing, Ready, Failed)
  • AGE: Time since agent creation

Creating a New AI Agent

Initialize a new AI agent with specific configuration:

kubectl faros ai-agents init \
  --name <agent-name> \
  --backend openai \
  --model gpt-4 \
  --api-key <your-api-key>

Configuration Options

  • --name: Name of the AI agent
  • --backend: AI backend provider (default: openai)
  • --model: Model to use (e.g., gpt-4, gpt-3.5-turbo)
  • --api-key: API key for the backend (creates a Kubernetes secret)
  • --secret-name: Use an existing Kubernetes secret instead of creating one
  • --secret-key: Key within the secret containing the API key
  • --namespace or -n: Kubernetes namespace for the agent (default: current namespace)

Using Existing Secrets

If you prefer to manage secrets separately:

# First, create a secret with your API key
kubectl create secret generic openai-credentials \
  --from-literal=api-key=<your-api-key>

# Then reference it when creating the agent
kubectl faros ai-agents init \
  --name my-agent \
  --backend openai \
  --model gpt-4 \
  --secret-name openai-credentials \
  --secret-key api-key

Agent Lifecycle

When you create an agent, the CLI:

  1. Creates an Agent resource in the intelligence.faros.sh/v1alpha1 API group
  2. Creates or references a Kubernetes secret for authentication
  3. Waits for the agent to transition to Ready phase
  4. Displays status and next steps

Example Workflow

# List existing agents
kubectl faros ai-agents list

# Create a new agent for production cluster analysis
kubectl faros ai-agents init \
  --name prod-analyzer \
  --backend openai \
  --model gpt-4 \
  --api-key sk-...

# Wait for agent to be ready
# Output: Agent "prod-analyzer" created and is now Ready

# List agents to verify
kubectl faros ai-agents list

Backend Support

Currently supported AI backends:

  • OpenAI: GPT-4, GPT-3.5-turbo, and other OpenAI models

Additional backends may be supported in future releases.