anomaly-detection

This Skill detects statistical anomalies in CloudWatch and Prometheus metrics by executing Python scripts to identify deviations. It serves as an input source for incident response systems to act on detected metric anomalies.

You can automate incident response by integrating detected statistical anomalies in CloudWatch and Prometheus metrics with your response systems. The Skill processes metric data using Python libraries to provide actionable insights for quick issue resolution.

Running anomaly detection requires access to CloudWatch and Prometheus metrics, the ability to execute Python scripts, and the awslabs.cloudwatch-mcp-server and awslabs.prometheus-mcp-server dependencies configured in your environment.

You monitor error rate metrics for statistical anomalies by applying the anomaly detection Skill to specific services like 'rag-qa-agent'. It analyzes the metric data using Python libraries to automatically identify statistical deviations.

The anomaly detection works with both CloudWatch and Prometheus simultaneously. It integrates with both platforms through their respective MCP servers to detect statistical anomalies across your infrastructure metrics for incident response.