Namespace LMKit.Agents.Observability

Collects and exposes metrics about agent execution.

Tracks counts, durations, and other statistics that can be exported to monitoring systems.

Represents a traced execution span for agent operations.

Spans track the duration, status, and metadata of agent operations, forming a hierarchical trace of execution.

Provides global configuration for agent tracing.

Configure the default tracer here to enable tracing across all agent executions.

A tracer that forwards to multiple underlying tracers.

Use this to send traces to multiple destinations simultaneously (e.g., console for debugging and OpenTelemetry for production monitoring).

A tracer that outputs to the console for debugging.

Useful during development to see agent execution flow. Not recommended for production use.

A tracer that stores spans in memory for inspection.

Useful for testing, debugging, and building custom trace visualizations.

A tracer implementation that does nothing.

Use this when tracing is disabled or not needed. All operations return immediately with minimal overhead.

An agent executor wrapper that adds tracing to all operations.

Wraps an existing executor and adds spans for agent execution, tool calls, planning steps, and other operations.

Interface for tracing agent operations.

Implement this interface to integrate with your preferred observability system (OpenTelemetry, Application Insights, custom logging, etc.).