Skill UI

This comprehensive runbook guides incident response for Groq API failures. It covers triage steps, severity classification (P1-P4), troubleshooting decision trees (e.g., handling 429 errors or 5xx outages), and implementing immediate mitigations like model fallback or key rotation. Use this when responding to Groq outages to restore service stability.

This skill provides comprehensive observability tools for Groq API integrations. It instruments the client to track critical performance metrics such as Time to First Token (TTFT), total latency, token throughput, and estimated cost. Furthermore, it sets up Prometheus exporters and alert rules to monitor rate limit status and detect performance degradation in real-time, ensuring robust production monitoring.