Amazon Bedrock AgentCore allows developers to create and deploy generative AI agents easily, offering a serverless runtime that integrates large language models with various data sources and tools. However, monitoring these agents can be challenging due to the complexity of the systems involved. The article explains how to use Grafana Cloud, along with OpenTelemetry and Amazon CloudWatch, to gain visibility into the performance and resource usage of these agents, which is critical for effective debugging and cost management. Grafana Cloud provides a unified platform that aggregates metrics from multiple AWS accounts and regions, eliminating data silos. This centralized monitoring allows engineers to track essential metrics like latency, token usage, and execution reliability across their AI infrastructure. The Cloud Provider Observability application in Grafana Cloud offers pre-built dashboards and alerts tailored for AWS resources, enabling instant access to key performance indicators. This saves time and simplifies the process of monitoring, allowing users to quickly assess the health of their AI agents. To get started, the article outlines steps to configure CloudWatch metric streams for Bedrock AgentCore agents. The process includes creating a Firehose stream in Amazon Data Firehose, setting it up in CloudWatch, and then linking it to Grafana Cloud. Once configured, data flows seamlessly into Grafana dashboards, providing a comprehensive view of the AI agents' performance. Users can filter metrics by various parameters, making it easier to troubleshoot and optimize their operations.