Skip to content

OpenTelemetry Span Event Not Recording Fix

DodaTech Updated 2026-06-24 3 min read

In this tutorial, you'll learn about OpenTelemetry Span Event Not Recording Fix. We cover key concepts, practical examples, and best practices.

Your OpenTelemetry span events do not appear in the tracing backend — span.add_event() calls produce no visible output, or the events lack timestamps and attributes. Events must be added to active spans and exported correctly.

The Problem

from opentelemetry import trace

tracer = trace.get_tracer(__name__)

def process_file(filename):
    with tracer.start_as_current_span("process-file") as span:
        # WRONG: event name as keyword argument
        span.add_event(name="file.opened")

The event is created but may not appear in the backend because name is a positional parameter, not a keyword argument. Also, events without attributes provide limited debugging value.

Step-by-Step Fix

1. Add events with attributes

from opentelemetry import trace

tracer = trace.get_tracer(__name__)

def process_file(filename):
    with tracer.start_as_current_span("process-file") as span:
        span.add_event("file.opened", {
            "filename": filename,
            "size_bytes": 1024,
        })
        # Process file...
        span.add_event("file.processed", {
            "filename": filename,
            "records": 500,
        })

2. Record exceptions as events

from opentelemetry import trace

tracer = trace.get_tracer(__name__)

def process_data(data):
    with tracer.start_as_current_span("process-data") as span:
        try:
            result = transform(data)
            span.set_attribute("data.valid", True)
        except ValueError as e:
            span.set_attribute("data.valid", False)
            span.record_exception(e, {
                "data.size": len(data),
                "data.preview": str(data)[:100],
            })

3. Use semantic convention event names

from opentelemetry import trace
from opentelemetry.semconv.trace import SpanAttributes

tracer = trace.get_tracer(__name__)

def cache_lookup(key):
    with tracer.start_as_current_span("cache-lookup") as span:
        span.add_event("cache.miss", {
            "cache.key": key,
        })
        # Fetch from database...
        span.add_event("db.query", {
            "db.system": "postgresql",
            "db.statement": "SELECT * FROM users WHERE id = ?",
        })

4. Add timed events

import time
from opentelemetry import trace

tracer = trace.get_tracer(__name__)

def deploy_service():
    with tracer.start_as_current_span("deploy") as span:
        span.add_event("deploy.started")

        time.sleep(2)
        span.add_event("deploy.built", {
            "image": "myapp:latest",
            "size_mb": 245,
        })

        time.sleep(3)
        span.add_event("deploy.deployed", {
            "environment": "staging",
            "replicas": 3,
        })

        time.sleep(1)
        span.add_event("deploy.completed")

5. Query events with TraceQL

from opentelemetry import trace

tracer = trace.get_tracer(__name__)

def payment_flow():
    with tracer.start_as_current_span("payment-flow") as span:
        span.add_event("payment.authorized", {
            "amount": 99.99,
            "currency": "USD",
        })
        span.add_event("payment.captured", {
            "transaction_id": "txn-123",
        })

Expected output:

process-file
  Events:
    file.opened @ 10:00:00.000
      filename: "data.csv"
      size_bytes: 1024
    file.processed @ 10:00:00.234
      filename: "data.csv"
      records: 500

Prevention Tips

  • Always add attributes to events for debugging value
  • Use span.record_exception(e) instead of manual exception events
  • Use semantic convention names for consistency
  • Add events at meaningful points in the operation lifecycle
  • Keep event names concise and namespaced (e.g., cache.miss, db.query)

Common Mistakes with span event

  1. Using return to exit a function early instead of wrapping a pure value in the monad
  2. Mixing let bindings with <- bindings in do notation, producing type errors
  3. Overlapping type class instances that cause GHC to reject the program with ambiguous dispatch errors

These mistakes appear frequently in real-world OPENTELEMETRY code. DodaTech's contributors have identified these patterns through analysis of open-source projects and production systems.

Practice Exercise

Write a pure function that safely divides two integers using Maybe, then test it with edge cases like division by zero and negative numbers.

This exercise reinforces the concepts covered in this guide. Try implementing it before checking online solutions.

FAQ

### What's the difference between span events and logs?

Span events are attached to a specific span and share its trace context. Logs are independent records that can be correlated via trace ID. Use events for lifecycle milestones within an operation, logs for general diagnostic output.

How many events can a span have?

There is no hard limit, but each event adds to the span size. For high-throughput spans, limit events to 5-10 per span. Use the batch span processor to manage export size. Events with many attributes increase storage and transfer costs.

Can I search for events in Grafana Tempo?

Yes, TraceQL supports event filtering: { span.events.name = "cache.miss" }. You can filter by event name and event attributes. This makes events searchable and actionable in your observability platform.

Built by the developers of DodaTech

Doda Browser, DodaZIP & Durga Antivirus Pro