Getting Started with AI Agent Observability using Gemini Enterprise Agent Platform and Cloud Trace

Gemini Enterprise Agent Platform provides tools for deploying AI agents, with Agent Garden supplying prebuilt templates and Agent Runtime handling their execution. These agents generate logs, metrics, and usage data that can be monitored to understand performance, reliability, and cost.

Monitoring these agents with services like Google Logs Explorer and Google Cloud Monitoring allows teams to evaluate reliability, detect performance issues, and understand cost drivers in production environments. By analyzing structured logs, tracking latency and request metrics, and configuring alerts, teams can build feedback loops that continuously improve agent behavior and efficiency.

In this lab, you will deploy a travel concierge agent, generate and analyze logs in Log Explorer, and create metrics and alerting policies in Cloud Monitoring

Learning objectives

Upon completion of this beginner-level lab, you will be able to:

• Deploy a pre-built AI travel agent on Gemini Enterprise Agent Platform
• Observe and interpret dashboards and logs generated by the agent in Vertex AI Agent Runtime
• Inspect the agent's traces and sessions to understand its decision-making process and identify potential issues

Intended audience

• DevOps Engineers​
• FinOps Practitioners​
• ML Engineers managing AI workloads

Prerequisites

Familiarity with the following will be beneficial but is not required:

• Basic understanding of AI agent concepts
• Google Cloud Monitoring

The following content can be used to fulfill the prerequisites:

• Monitor Compute Engine Resources Through Cloud Monitoring
• Introduction to Agentic AI on Google Cloud

Updates

June 15th, 2026 - Updated lab to reflect the incorporating of Vertex AI into the new Gemini Enterprise Agent Platform