KubeRadar

KubeRadar is a Kubernetes debugging and diagnostics CLI that helps developers diagnose and debug issues in clusters.

Installation

go install github.com/Vemula-Rohith/kuberadar/cmd/kuberadar@latest

Or build from source:

git clone https://github.com/Vemula-Rohith/kuberadar
cd kuberadar
go build -o kuberadar ./cmd/kuberadar

Usage

Commands

• kuberadar sweep — Sweep all pods in a namespace for issues
• kuberadar pod [name] — Diagnose a specific pod or all pods in namespace (use --diagnose for full evidence and recommendation)
• kuberadar deployment [name] — Diagnose a deployment by checking its pods
• kuberadar explain [issue-id] — Explain an issue ID and how to resolve it (self-documenting)
• kuberadar version — Print the version

Flags

• --kubeconfig — Path to kubeconfig file
• -n, --namespace — Kubernetes namespace (default from kubeconfig context)
• -o, --output — Output format: table (default), json

Examples

# Sweep all pods in default namespace
kuberadar sweep

# Sweep pods in a specific namespace
kuberadar sweep -n my-namespace

# Diagnose a specific pod (summary)
kuberadar pod my-pod -n my-namespace

# Full diagnosis with evidence and recommendation
kuberadar pod my-pod -n my-namespace --diagnose

# Output as JSON for CI
kuberadar sweep -o json

# Explain an issue ID
kuberadar explain KR003

Exit Codes (CI integration)

Code	Meaning
0	No issues
1	Warnings only
2	Critical issues

Example: kuberadar sweep || exit 1 fails the pipeline on any issues.

Issue IDs

ID	Description	Severity
KR001	OOMKilled	Critical
KR002	CrashLoopBackOff	Critical
KR003	ImagePullBackOff	Critical
KR004	UnschedulablePod	Warning

Architecture

• CLI — Cobra commands and flags
• Engine — Orchestrates diagnostics via Run(scope)
• Detectors — Problem detection logic (crashloop, OOM, image pull, scheduling)
• Providers — Thin client-go wrappers for Kubernetes API
• Context Builder — Builds context per resource (events, node)