Kubectl Debugging

What it does

This skill enables AI agents to execute kubectl commands directly within Kubernetes clusters for troubleshooting and inspection. It allows agents to retrieve detailed pod status, inspect container logs, and view resource definitions without manual CLI intervention.

When to use it

• Diagnosing why a specific pod is stuck in CrashLoopBackOff or Pending states.
• Retrieving recent error logs from containers to identify application failures.
• Verifying the current configuration of Deployments, Services, or ConfigMaps during deployment issues.
• Checking node resource utilization when cluster performance degrades unexpectedly.

Key capabilities

• Executes standard kubectl subcommands like get, describe, logs, and exec.
• Filters output to focus on specific namespaces, labels, or pod names.
• Interprets complex Kubernetes status messages into actionable insights.
• Supports context switching between different cluster configurations.

Example prompts

• "Describe the 'production-frontend' pod in the 'prod-us-east' namespace and explain why it is not running."
• "Show me the last 50 lines of error logs for any container named 'api-server' across all namespaces."
• "Get the YAML definition for the 'database-service' and check if the port mapping matches the deployment spec."

Tips & gotchas

Ensure your AI agent has valid credentials and access to the target Kubernetes cluster context before running diagnostic commands. Be mindful of rate limits on large clusters when requesting logs from many pods simultaneously, as this can trigger API throttling.