Skip to main content

Observability Tracing

Distributed tracing system for monitoring and debugging RecoAgent applications.

Overview​

The observability tracing system provides comprehensive distributed tracing capabilities for monitoring request flows, performance bottlenecks, and system behavior.

Core Features​

  • Distributed Tracing: Track requests across multiple services
  • Performance Monitoring: Monitor latency and throughput
  • Error Tracking: Track and analyze errors
  • Integration: LangSmith and custom tracing support

Usage Examples​

Basic Tracing Setup​

from recoagent.observability.tracing import TracingManager

# Create tracing manager
tracing_manager = TracingManager()

# Start trace
with tracing_manager.start_trace("user_query_processing") as trace:
    # Process query
    result = process_query("What is machine learning?")
    trace.add_metadata({"query_length": len("What is machine learning?")})

Advanced Tracing​

# Create span for specific operation
with tracing_manager.create_span("rag_retrieval") as span:
    # Perform retrieval
    documents = retrieve_documents(query)
    span.add_attributes({
        "documents_found": len(documents),
        "retrieval_time": time.time() - start_time
    })

API Reference​

TracingManager Methods​

start_trace(name: str) -> Trace​

Start a new trace

Parameters:

  • name (str): Trace name

Returns: Trace context manager

create_span(name: str) -> Span​

Create a new span

Parameters:

  • name (str): Span name

Returns: Span context manager

See Also​