kube-medic
par tkuehnl
Coding Agents & IDEsRésumé
Kubernetes Cluster Triage & Diagnostics — instant AI-powered incident triage via kubectl.
Description
Ce que fait ce skill
kube-medic runs Kubernetes diagnostics through kubectl. It covers cluster health sweeps, pod autopsies, deployment analysis, resource pressure detection, and event monitoring. Data across logs, events, and resource metrics is correlated to surface root causes, and write operations require explicit user confirmation before execution.
Instead of running separate kubectl commands for logs, events, and resource metrics, it correlates all three in a single triage workflow to produce actionable diagnoses rather than a list of symptoms.
Quand l'utiliser
- Diagnosing a CrashLoopBackOff pod during an on-call incident
- Checking cluster health before a production deployment
- Finding which pods are consuming the most memory on overloaded nodes
- Tracing recent warning events after an unexpected outage
- Rolling back a broken deployment after identifying a bad image
Exemple de workflow
User asks: Something is wrong with my cluster, can you check?
- 1
Runs sweep subcommand to get a cluster-wide health overview
- 2
Identifies CrashLoopBackOff pods and NotReady nodes from sweep output
- 3
Runs pod autopsy on the most suspicious pod to retrieve logs, events, and container status
- 4
Runs resources subcommand to check for memory or CPU pressure contributing to the issue
- 5
Correlates findings and presents a diagnosis with specific kubectl commands awaiting user approval
Triage report showing root cause (e.g., OOMKilled due to low memory limit), a timeline of related events, and a rollback or scaling command the user can confirm to execute
Installation
clawhub install tkuehnl/kube-medicExécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.