OpenTelemetry Instrumentation for Django

Warnings

• breaking As of version 0.61b0, support for Django versions older than 2.0 has been dropped. Applications using older Django versions will need to upgrade Django or pin to an older `opentelemetry-instrumentation-django` version.
  Fix: Upgrade Django to version 2.0 or newer, or use `opentelemetry-instrumentation-django<0.61b0`.
• gotcha The library is in a beta (`b0`) release stage, indicating potential API changes or instability. While widely used, be aware of this when deploying to production and monitor for updates.
  Fix: Stay updated with releases, review changelogs, and conduct thorough testing before deploying new beta versions.
• gotcha When using `request_hook` to add attributes to spans, attributes dependent on Django's middleware (e.g., `request.user`, `request.site`) will not be available. These should be attached in the `response_hook` instead, which executes after all middlewares have run.
  Fix: Place logic for adding middleware-dependent attributes into the `response_hook` callback of `DjangoInstrumentor().instrument()`.
• gotcha For comprehensive automatic instrumentation (e.g., database drivers like psycopg2, redis, requests), the `opentelemetry-bootstrap -a install` command should be run in your environment. This command scans installed packages and installs appropriate instrumentation libraries.
  Fix: Execute `opentelemetry-bootstrap -a install` after installing `opentelemetry-distro` and `opentelemetry-instrumentation-django` to ensure all detectable components are instrumented.

Install

• pip install opentelemetry-instrumentation-django opentelemetry-distro opentelemetry-exporter-otlp Full instrumentation setup

Imports

• DjangoInstrumentor

  from opentelemetry.instrumentation.django import DjangoInstrumentor

Quickstart

This quickstart demonstrates how to manually initialize OpenTelemetry and instrument a Django application. The `DjangoInstrumentor().instrument()` call sets up automatic tracing for HTTP requests, database queries, and other Django components. For a zero-code setup, `opentelemetry-instrument` wrapper is commonly used (e.g., `opentelemetry-instrument python manage.py runserver`).

import os
from opentelemetry import trace
from opentelemetry.sdk.resources import Resource
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor, ConsoleSpanExporter
from opentelemetry.instrumentation.django import DjangoInstrumentor

# Configure OpenTelemetry SDK
resource = Resource.create({"service.name": os.environ.get('OTEL_SERVICE_NAME', 'my-django-app')})
provider = TracerProvider(resource=resource)

# For demonstration, export traces to console
# In a real application, you'd use OTLPSpanExporter or similar
processor = BatchSpanProcessor(ConsoleSpanExporter())
provider.add_span_processor(processor)
trace.set_tracer_provider(provider)

# Instrument Django
DjangoInstrumentor().instrument()

print("OpenTelemetry Django instrumentation initialized.")
# This code snippet would typically be placed in a Django app's __init__.py 
# or a settings-loaded module, ensuring it runs during application startup.
# For a zero-code approach, use `opentelemetry-instrument python manage.py runserver`