The Intelligent Kubernetes Inspector
Not just monitoring โ understanding. Not just alerts โ answers. Not just problems โ solutions.
๐ Read the full documentation here
helm repo add k8s-inspector https://charts.k8s-inspector.dev
helm install k8s-inspector k8s-inspector/k8s-inspector --namespace k8s-inspector --create-namespace
kubectl port-forward svc/k8s-inspector-frontend 3000:80 -n k8s-inspector
git clone https://github.com/egkristi/k8s-inspector.git
cd k8s-inspector
docker-compose -f deploy/local/docker-compose.yml up -d
Access:
Not just โPod is crashingโ โ but why, and how to fix it.
Real savings recommendations, automated rightsizing (GitOps PRs), and eBPF-powered carbon/power tracking.
Opt-in LLM integration to explain complex failures (OOMKilled, CrashLoops) or cost spikes in plain English.
Fix problems before they happen with ML-powered anomaly detection.
CIS benchmarks, NSA/CISA guidelines โ with one-click remediation.
One dashboard, unlimited clusters.
The Kubernetes tooling ecosystem is fragmented. You usually need one tool for cost, one for security, and another for monitoring. k8s-inspector unifies them and adds automated remediation.
| Capability | ๐ฆโโฌ k8s-inspector | Kubecost | Datadog | Robusta | Lens / K9s |
|---|---|---|---|---|---|
| Cost & Waste Analysis | โ Yes | โ Yes | โ Yes ($$) | โ No | โ No |
| GreenOps (Watt/CO2) | โ Yes (eBPF) | โ No | โ No | โ No | โ No |
| Security Auditing | โ Yes | โ No | โ Yes | โ No | โ No |
| Auto-Remediation | โ Yes (Playbooks) | โ No | โ No | โ Yes | โ No |
| GitOps Auto-PRs | โ Yes | โ No | โ No | โ No | โ No |
| Predictive ML | โ Yes | โ No | โ Yes | โ No | โ No |
| Agentic AI Chat | โ Yes (Opt-in) | โ No | โ Yes (SaaS) | โ Yes | โ No |
| Delivery Model | Open Source (AGPLv3) | Open Core | SaaS Only | Open Source | Open Core |
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ k8s-inspector 2.0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Next.js 14 Dashboard (WebSocket Real-time) โ
โ โ โ
โ FastAPI Backend (Async Python) โ
โ โ โ
โ PostgreSQL + TimescaleDB (Metrics) โ
โ Redis (Cache + Pub/Sub) โ
โ โ โ
โ Kubernetes Clusters (N) โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
Cluster Health: Control plane, etcd, nodes, operators, certificates
Workloads: Pods, deployments, StatefulSets, restarts, resources
Cost: Compute, storage, network โ with optimization tips
Security: CIS benchmarks, pod security, network policies
Performance: Anomaly detection, predictions, capacity planning
| Component | Technology |
|---|---|
| Frontend | Next.js 14 + Tailwind + shadcn/ui |
| Backend | FastAPI (Python async) |
| Database | PostgreSQL + TimescaleDB |
| Cache | Redis |
| ML | scikit-learn + PyTorch |
| Deploy | Helm + Docker |
See CONTRIBUTING.md for guidelines.
k8s-inspector uses a dual-license model:
See LICENSING.md for full details, features, and terms.