| name | k8s-healthcheck |
| description | Diagnose Kubernetes workload health with structured reports and actionable remediation steps. |
Kubernetes Health Check Skill
Description
Diagnose the health of Kubernetes workloads by inspecting pod status, events, logs, resource usage, and node conditions. Produces a structured health report with actionable remediation steps.
Triggers
- User asks to "check health", "diagnose", or "troubleshoot" a workload
- User reports pods are crashing, not starting, or misbehaving
- User asks "why is X not working" for a Kubernetes resource
Instructions
Health check sequence
- Get the target resource (Deployment, StatefulSet, DaemonSet, or Pod) via
k8s_get_resources
- List all pods owned by the resource and their current status
- For any pod not in Running/Ready state, collect:
- Pod details via
k8s_describe_resource
- Container logs via
k8s_get_pod_logs (last 50 lines, including previous container if restarted)
- Check for recent events in the namespace via
k8s_get_events
Diagnosis rules
- Pending + no events: Likely scheduling issue — check node resources, taints, affinity
- Pending + FailedScheduling: Report which constraint failed (CPU, memory, node selector)
- ImagePullBackOff: Check image name, tag, and registry auth
- CrashLoopBackOff: Read logs from previous container, check OOMKilled
- Running but not Ready: Check readiness probe config and endpoint
- Evicted: Check node disk/memory pressure
Report format
Produce a structured report:
RESOURCE: <name> in <namespace>
STATUS: <Healthy|Degraded|Unhealthy>
PODS: <ready>/<total> ready
ISSUES:
- [SEVERITY] <issue description>
EVIDENCE: <what you found>
FIX: <recommended action>
Validation Criteria
- All relevant pods are inspected
- Root cause is identified (not just symptoms)
- Remediation steps are specific and actionable
- No false positives (don't flag healthy resources)
Example
Input: "Check the health of the nginx deployment in namespace demo"
Expected tool sequence:
- k8s_get_resources (get deployment)
- k8s_get_resources (list pods)
- k8s_describe_resource (describe unhealthy pods)
- k8s_get_pod_logs (get container logs)
- k8s_get_events (check for error events)
Expected output: "RESOURCE: nginx in demo | STATUS: Healthy | PODS: 3/3 ready | No issues detected."
