Observability6 min read
Agent observability: traces that show what the agent did, not only latency
When something goes wrong, latency dashboards cannot tell you which tool call or plan step caused the damage.
- Trace the plan, each tool call, handoffs, and final outcome under one run ID.
- Separate performance metrics from quality evals; both are required.
- Attribute cost to workflows and roles, not only to model providers.