This project provides a ChatGPT bot for Kubernetes issues, specifically designed to help users resolve Prometheus alerts. It targets Kubernetes operators and DevOps engineers seeking AI-powered assistance for incident response, offering quick, AI-generated solutions to common alerts.

How It Works

The bot leverages the Robusta.dev platform to receive Prometheus alerts via a webhook. It then queries OpenAI's GPT-3.5-turbo model (referred to as ChatGPT) to generate explanations and potential solutions for the alerts. This approach aims to provide immediate, actionable insights without manual research, integrating directly into the alert response workflow.

Quick Start & Requirements

• Install: Install Robusta using Helm and load the ChatGPT playbook.
• Prerequisites: A Slack workspace and an OpenAI API key.
• Configuration: Add the ChatGPT playbook repository URL and your OpenAI API key to generated_values.yaml.
• Deployment: Run helm upgrade robusta robusta/robusta --values=generated_values.yaml --set clusterName=<YOUR_CLUSTER_NAME>.
• Alerting: Configure Prometheus to send alerts to Robusta.
• Demo: Deploy a sample pending pod and trigger a KubePodCrashLooping alert to test the Slack integration.
• Resources: Robusta.dev

Highlighted Details

• Integrates directly with Prometheus alerts via webhook.
• Provides an "Ask ChatGPT" button on alerts within Slack.
• Built on the Robusta.dev open-source platform for Kubernetes alert response.

Maintenance & Community

The project is being deprecated and replaced by HolmesGPT. Community discussion and feature suggestions are encouraged on their Slack channel.

Licensing & Compatibility

The README does not explicitly state the license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project is marked for deprecation. The underlying model used is GPT-3.5-based (text-davinci-003), not the exact ChatGPT model. Future improvements are suggested to include pod logs and kubectl get events for richer context.