thndrai/thndr repository overview

⁠ThndrAI Agent

⁠Proactive Kubernetes Monitoring & AI-Powered Analysis

The ThndrAI Agent is a proactive monitoring and analysis tool designed to maintain the health and stability of your Kubernetes clusters. It operates as an autonomous agent, continuously observing resources, identifying issues using AI-driven root cause analysis, and providing on-demand reporting to help your team resolve incidents faster.

⁠Features

The agent is equipped with a suite of powerful tools to manage, diagnose, and report on your Kubernetes environment. Key capabilities include:

• Advanced Incident Reporting: Automatically generate and email detailed incident reports tailored to different audiences (e.g., Engineers, Product Managers) with AI-powered summaries.

• Incident Querying & Management: List, filter, and get detailed views of all incidents. You can filter by severity, status, namespace, and resource type.

• De-duplicated Incident Summaries: Get a clean, de-duplicated summary showing only the first occurrence of each unique issue to reduce noise.

• Cost & Resource Analysis: Analyze the estimated resource cost of any namespace and get AI-driven suggestions to optimize CPU and memory for your deployments.

• Security Scanning: Scan container images for Common Vulnerabilities and Exposures (CVEs) based on a defined severity threshold.

• Network Diagnostics: Perform network traces and connectivity checks between pods to diagnose issues with NetworkPolicies, Services, and DNS.

• Cluster & CI/CD Management: Trigger on-demand cluster backups, manage EKS cluster lifecycles (e.g., version updates), and interact with GitHub Actions to re-run failed jobs.

⁠Changelogs

⁠[1.0] - 2025-09-22

This is the first public release of the ThndrAI Agent.

⁠Added

• Initial support for Kubernetes v1.25 and newer on EKS, GKE, and AKS.
• Local-first architecture: agent and database run entirely in the user's environment.
• AI-powered analysis using Google Gemini models via a user-provided API key.
• Conversational interface through the VS Code Copilot extension.
• Proactive detection of common issues like CrashLoopBackOff, OOMKilled, and resource misconfigurations.
• License key validation with the ThndrAI licensing server.