Building AI-Assisted Operations: Agentic AI Workshop

Welcome to the Agentic AI Workshop at CfgMgmtCamp Ghent 2026.

In this hands-on workshop, you’ll learn how to build AI-assisted operational workflows using Kubernetes-native AI agents. By the end of this workshop, you’ll have a multi-agent system that can autonomously investigate and fix infrastructure issues.

What You’ll Build

graph TD
    A[Orchestrator Agent<br/>Coordinates investigation and remediation] --> B[K8s Agent<br/>Diagnose]
    A --> C[PromQL Agent<br/>Metrics]
    A --> D[Pulumi Agent<br/>Fix PRs]

    B --> E[kubectl<br/>MCP Server]
    C --> F[Prometheus<br/>MCP Server]
    D --> G[GitHub<br/>MCP Server]

    style A fill:#6366f1,stroke:#4f46e5,color:#fff
    style B fill:#22c55e,stroke:#16a34a,color:#fff
    style C fill:#22c55e,stroke:#16a34a,color:#fff
    style D fill:#22c55e,stroke:#16a34a,color:#fff
    style E fill:#f59e0b,stroke:#d97706,color:#fff
    style F fill:#f59e0b,stroke:#d97706,color:#fff
    style G fill:#f59e0b,stroke:#d97706,color:#fff

Your multi-agent system will diagnose Kubernetes issues (pods, deployments, events), analyze metrics and resource utilization, and create pull requests to fix infrastructure code.

Workshop Chapters

Chapter	Title	Duration
01	Create Kubernetes Cluster	30 min
02	Deploy Kagent & MCP	45 min
03	Deploy Platform Workload	30 min
04	Multi-Agent Troubleshooting	60 min
05	Cleanup	15 min

Prerequisites

Before starting, make sure you have:

A GitHub account with a personal access token
A Pulumi account (free tier is sufficient)
Basic understanding of Kubernetes concepts
Familiarity with TypeScript or YAML

Getting Started

Option 1: GitHub Codespaces (Recommended)

Click the button below to open this repository in GitHub Codespaces with all tools pre-installed:

Option 2: Local Development

Clone the repository:

git clone https://github.com/dirien/cfgmgmtcamp-2026-agentic-ai-workshop.git
cd cfgmgmtcamp-2026-agentic-ai-workshop

Install the required tools:
- Pulumi CLI
- kubectl
- Node.js (LTS version)
- doctl (DigitalOcean CLI - optional)
Login to Pulumi:
```
pulumi login
```

Workshop Credentials

Your instructor will provide:

LLM Endpoint: Pre-configured DigitalOcean GenAI endpoint
LLM API Key: Shared API key for the workshop (via Pulumi ESC)

These credentials are automatically available when you use the workshop’s Pulumi ESC environment.

Tips for Success

Follow the chapters in order - Each chapter builds on the previous one
Don’t skip the verification steps - They help catch issues early
Ask questions! - The instructors are here to help
Experiment - Try modifying the agents and see what happens

Resources

Let’s start with Chapter 1: Create Kubernetes Cluster.