About
The Observability Alert Manager provides a robust framework for monitoring Claude Code performance and behavior using Grafana and Loki. It allows developers to define, list, test, and manage automated alerts for specific events such as high error rates, context window saturation, long session durations, and unusual subagent spawning. By integrating directly with enhanced telemetry data, it ensures that teams are immediately notified via Slack, email, or PagerDuty when Claude Code encounters operational anomalies or exceeds predefined resource thresholds, helping maintain optimal AI performance and cost control.
