Effective alert management is essential for quickly identifying problems and maintaining operational resilience. The article explains how to manage alerting rules in Amazon Managed Service for Prometheus (AMP), particularly for a hypothetical company, Example Corp, that monitors an Elastic Kubernetes Service (EKS) workload. The focus is on creating, routing, and administering alerting rules to avoid alert fatigue while ensuring the rules are understandable and actionable. Example Corp collects metrics from multiple AWS accounts using a centralized observability architecture. It relies on recording rules for optimizing query performance by precomputing complex Prometheus Query Language (PromQL) expressions. This allows the team to create new metrics that serve as the foundation for alerting rules. Alerting rules themselves are designed to detect incidents proactively, triggering notifications when specific conditions, like high CPU usage or HTTP error rates, are met. For instance, the article provides a practical example where Example Corp defines a recording rule to calculate the application error rate based on HTTP requests. It divides the total number of 500 status errors by the total number of requests over a five-minute window. An alert is then set to notify the on-call team if the error rate exceeds 10% for more than five minutes. This setup streamlines the monitoring process, enhances performance, and simplifies the response to incidents.