env-troubleshoot

name	env-troubleshoot
metadata	{"author":"rhoai3-demo","version":"2.0.0","platform-family":"rhoai","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"Demo Environment"}
description	Diagnose and fix issues in the live rhoai3-demo AWS/OpenShift environment once active deploy and validate scripts exist. During the reimplementation, use this skill to rebuild troubleshooting coverage from legacy references. Use when a deployment stage fails, validate.sh reports errors, pods are in CrashLoopBackOff/Pending/Error/ImagePullBackOff, ArgoCD shows OutOfSync or Degraded, operators are not installing, GPU nodes are not joining, InferenceServices are not Ready, LlamaStack returns errors, Guardrails Orchestrator is unhealthy, MCP tools are failing, or the user reports any problem with their OpenShift AI environment. Also use when the user asks "why isn't X working?" for any demo component. Do NOT use for chatbot UI/prompt changes (use rhoai-chatbot-customization) or model evaluation workflows (use rhoai-model-evaluation).

RHOAI Troubleshooting

Structured diagnostic workflow for resolving issues with the RHOAI demo on the active product baseline in docs/PLATFORM_BASELINE.md.

Reimplementation Status

The active implementation is being rewritten. No active deploy or validate scripts exist yet. Treat legacy step names, validation commands, and diagnostic patterns in this skill as reference material for rebuilding troubleshooting coverage, not as active-project instructions.

Do not run scripts from backup/legacy-implementation-2026-06-09/ unless the user explicitly asks to restore or inspect the legacy implementation.

When to Use

• A validate.sh script reports failures
• A deploy.sh script errors out or hangs