Observability Overview

Observability ensures AI systems are reliable, safe, and high-quality by enabling continuous monitoring and evaluation. For technical leaders and developers, it is key to building trust and maintaining performance in generative AI applications.

With Azure AI Foundry, observability means:

• Using evaluators to measure quality, safety, and reliability at every stage—from model selection to production.
• Assessing models and applications before deployment, and monitoring them in real time after launch.
• Tracking metrics like relevance, groundedness, and safety to quickly identify and address risks.
• Leveraging integrated dashboards and Azure Monitor for actionable insights.

Adopting observability best practices helps organizations deliver trustworthy AI, reduce risk, and meet business goals.

For more details on observability, see the Azure AI Foundry Observability Overview.