k8s-diagnose

name	k8s-diagnose
description	Diagnose common Kubernetes issues: pod crashes, resource bottlenecks, scheduling failures, and networking problems.

K8s Diagnose

Workflow

1. Use pod_list or pod_get to identify failing pods
2. Use pod_logs to fetch recent logs from crashed containers
3. Use event_list filtered by Warning type to find relevant events
4. Use resource_get to check node conditions and resource capacity
5. Summarize findings and suggest remediation steps

Common Patterns

• Crashing pod → check pod_logs for application errors, verify image pull
• Pending pod → check event_list for scheduling failures (resources, taints)
• OOMKilled → check resource limits, suggest increasing memory
• ImagePullBackOff → verify image name, registry auth, network access