K8sGPT Auto Remediation

Automatically fix issues in your Kubernetes clusters with AI-powered remediation.

What is Auto Remediation?

Auto Remediation is an experimental feature that attempts to fix problems encountered in your Kubernetes clusters. It interprets K8sGPT results and applies patches to fix issues on target resources.

How it Works

1K8sGPT operator parses results and identifies resources that have auto remediation enabled
2Creates a Mutation resource to track the remediation process
3Calculates a patch based on similarity score with the original resource
4Applies the patch to fix the issue
5Watches the resource until the issue is resolved or persists

Supported Resources

Currently in Alpha state, the following Kubernetes resources are supported:

•Service
•Pod (Owned by ReplicaSet/Deployment)
•Pod (Static)

Configuration

To enable auto remediation, you need to configure the K8sGPT custom resource with the following fields:

apiVersion: core.k8sgpt.ai/v1alpha1
kind: K8sGPT
metadata:
  name: k8sgpt-sample
  namespace: default
spec:
  ai:
    autoRemediation:
      enabled: true
      riskThreshold: 90
      resources:
        - Pod
        - Service
        - Deployment

Mutations

Mutations are custom resources that hold the state and intent for mutating resources in the cluster. They will eventually be compatible with GitOps processes, allowing you to pull mutations out of cluster and re-apply them.

Currently, Mutations reside in the same namespaces as your K8sGPT custom resource and are controlled by a finalizer.

Important Note

This feature is highly experimental and is not ready for use in a production environment. Use with caution and always review the proposed changes before applying them.