Description
k8s-servant is a specialized tool designed to serve as a read-only AI Site Reliability Engineer (SRE) for Kubernetes clusters. Its primary function is to identify issues within a cluster, explain these issues with evidence, and provide guidance on how to resolve them, all without making any changes to the cluster itself. This ensures that users maintain full control over their environments.
Read More
The tool operates by mimicking the approach of an experienced cluster administrator. It answers questions such as "Are there unhealthy pods?" or "Why is X pending?" using read-only tools to gather evidence, which it then presents to the user. It also performs regular health checks, reporting on new, ongoing, or resolved issues, and provides a comprehensive morning report detailing the cluster's status.
Users can choose from a variety of models to power k8s-servant, including Microsoft Foundry, OpenAI, Anthropic, or a self-hosted model through Ollama. This flexibility allows users to select the model that best fits their needs and infrastructure. The tool's results can be outputted to a terminal or sent to an MS Teams channel for easy access and collaboration.
k8s-servant is built with Python and utilizes several libraries such as openai, anthropic, and azure-identity to facilitate its operations. It is designed to be secure, operating with read-only RBAC permissions and without access to Secrets. The tool is still in its early stages, having been tested on a limited number of clusters, so users are advised to verify its findings before taking action.
Overall, k8s-servant provides a valuable service for Kubernetes administrators, offering insights and recommendations while ensuring the safety and integrity of the cluster.
- CategoryAI DevOps & Cloud Tools
- Pricing ModelOpen Source
- DeploymentKubernetes
- PlatformsCloud
k8s-servant's Core Features
Read-only AI SRE for Kubernetes
Model choice: Microsoft Foundry, OpenAI, Anthropic, Ollama
Regular health checks and morning reports
Evidence-based issue explanations
Terminal and MS Teams output
Secure: no cluster changes, no Secrets access
Built with Python and asyncio
Supports multiple model providers
How to use k8s-servant?
Configure: Choose your preferred model
Deploy: Install k8s-servant in your cluster
Monitor: Receive health checks and reports
Analyze: Review findings and recommendations
k8s-servant's Use Cases
- Cluster Health Monitoring
- Issue Diagnosis
- Evidence-Based Reporting
- Model Flexibility
- Secure Operations
