AI engineering glossary · Agents tools MCP
What is agent observability?
Agent observability: The practice of recording and inspecting what an agent did, including each model call, tool call, input, output, latency and cost. Because agents take variable paths, this record is the main way to debug failures and measure quality.
Agent observability: an example
A dashboard shows that 12% of runs hit the step limit, most after repeated failures of the same search tool.
Related terms
More Agents tools MCP terms
- Agent skills
- Agent trace
- Agent-to-agent protocol
- Agentic workflow
- AI agent
- Approval gate
- AutoGen
- Autonomous agent
Where you learn it
Agent observability is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.