Diagnose a cloud outage from dashboard screenshots
When a Kubernetes cluster stopped scheduling pods, an infrastructure team fed screenshots to the agent, which guided them through the cloud console to IP-address exhaustion and gave the commands to fix it.
Steps
- Share screenshots of the failing dashboards
- Let it walk you through the console step by step
- Run the suggested commands yourself after reading them
Starter prompt
Our cluster stopped scheduling new pods 20 minutes ago. Here are screenshots of the node and pod dashboards. Walk me through where to look next in the console, one step at a time, and tell me the likely cause.
Guardrail: You run the commands; the agent explains them.
Prompts are our own examples. Results from sources are as reported by their authors.