> ## Documentation Index
> Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt
> Use this file to discover all available pages before exploring further.
# K8sGPT Features and Capabilities
> K8sGPT is an AI-driven tool for diagnosing and remediating issues in Kubernetes clusters, offering real-time insights and actionable advice.
Welcome to this lesson on **K8sGPT**, a CNCF Sandbox project that brings AI-driven diagnostics and remediation to your Kubernetes clusters. Developed by Alex Jones and his team, K8sGPT is fully open source and available as either a Kubernetes operator or a standalone binary. It connects to your live cluster, gathers resource state across Deployments, Services, ReplicaSets, and more, then leverages AI to pinpoint issues and recommend fixes.
## Key Features
| Feature | Description |
| ------------------------- | --------------------------------------------------------------------------- |
| Cluster Scanning | Crawls all namespaces and collects real-time status of Pods, Services, etc. |
| Issue Diagnosis | Identifies errors, misconfigurations, and failed probes across resources. |
| Plain-English Explanation | Translates technical findings into clear, human-readable summaries. |
| Actionable Advice | Provides step-by-step remediation steps to resolve detected issues. |
| Data Anonymization | Strips sensitive details before sending telemetry to the AI backend. |
| Plugin Architecture | Supports custom analyzers and integrations via an extensible plugin system. |
## Supported AI Backends
K8sGPT integrates seamlessly with major AI providers. You can choose from:
| AI Provider | Typical Use Case |
| -------------- | ------------------------------------------------------ |
| OpenAI | Access GPT-4, ChatGPT APIs |
| Azure OpenAI | Enterprise-grade LLMs hosted on Azure |
| Google Gemini | Google Cloud’s advanced generative AI models |
| Amazon Bedrock | AWS-managed large language models |
| Cohere | High-quality embeddings and language models |
| LocalAI | On-prem or local model hosting for stricter compliance |
Before running the analysis, verify your kubeconfig and context by running `kubectl get nodes`.\
Ensure you have network access to both your cluster and the configured AI backend.
## Quickstart: Cluster Analysis
After configuring your preferred AI provider, simply run:
```bash theme={null}
olama@bakugo:~/demo$ k8sgpt analyse --explain
100% AI Provider: openai
0: Deployment k8sgpt/k8sgpt-ollama
- Error: Deployment k8sgpt/k8sgpt-ollama has 1 replica but 0 are available.
Explanation: The Deployment expects one Pod, but none are running.
Solution:
1. Check Pod status: kubectl get pods -n k8sgpt
2. Describe the Pod: kubectl describe pod -n k8sgpt
3. Inspect logs: kubectl logs -n k8sgpt
4. Resolve errors in the Pod spec or logs.
1: Service hosting/web2
- Error: Service has no ready endpoints; expected 2 Pods but none are Ready.
Explanation: Pods backing this Service are not passing readiness probes.
Solution:
1. kubectl get pods -n hosting
2. kubectl logs -n hosting
3. Verify readinessProbe in the Pod spec.
4. Restart or fix the Pods until they become Ready.
2: Service k8sgpt/k8sgpt-ollama
- Error: No endpoints found for label app=k8sgpt-ollama.
Explanation: The Service selector isn’t matching any Pods.
Solution:
1. kubectl get pods -l app=k8sgpt-ollama -n k8sgpt
2. Label Pods if missing: kubectl label pod app=k8sgpt-ollama -n k8sgpt
3: ReplicaSet k8sgpt/k8sgpt-ollama-668396898
- Error: pods "k8sgpt-ollama-668396898-xxxxx" is forbidden: serviceaccount "k8sgpt" not found.
Explanation: The ReplicaSet can’t create Pods without a valid ServiceAccount.
Solution:
1. Create the ServiceAccount: kubectl create serviceaccount k8sgpt -n k8sgpt
2. Reapply or restart the Deployment/ReplicaSet.
```
This output demonstrates how K8sGPT scans your cluster, identifies common issues across Deployments, Services, and ReplicaSets, and provides clear, actionable remediation steps.
You will also see a live demo and hands-on lab with K8sGPT in the next module.
## Links and References
* [K8sGPT GitHub][k8sgpt-gh]
* [CNCF Sandbox][cncf-sandbox]
* [OpenAI API][openai-api]
* [Azure OpenAI][azure-openai]
* [Google Gemini][google-gemini]
* [Amazon Bedrock][amazon-bedrock]
* [Cohere][cohere]
* [LocalAI][localai]
* [Kubernetes Basics][k8s-basics]
[k8sgpt-gh]: https://github.com/alexjones/k8sgpt
[cncf-sandbox]: https://github.com/cncf/sandbox
[openai-api]: https://platform.openai.com/docs/api-reference
[azure-openai]: https://learn.microsoft.com/azure/cognitive-services/openai/
[google-gemini]: https://cloud.google.com/vertex-ai/generative-ai
[amazon-bedrock]: https://aws.amazon.com/bedrock/
[cohere]: https://cohere.com/
[localai]: https://github.com/go-skynet/LocalAI
[k8s-basics]: https://kubernetes.io/docs/concepts/overview/what-is-kubernetes/