Django ORM spans are missing while using Psycopg2 Instrumentator

[overide]

Hi guys,

I'm using Django 4.2 and instrumented the Django and Psycopg2 Instrumentor in the Gunicorn config file, while I can get everything else, Django ORM spans are not present in traces.

Note: I went through this discussion as well, but didn't help - open-telemetry/opentelemetry-python#2989

I'm using Grafana Tempo and Grafana Alloy for traces and Otel collection. Please find other details below -

Libraries I'm using

python 3.9.19
django = "==4.2"
psycopg2-binary = "==2.9.9"
opentelemetry-distro = {extras = ["otlp"], version = "^0.59b0"}
opentelemetry-exporter-otlp = "^1.38.0"
opentelemetry-instrumentation-django = "^0.59b0"
opentelemetry-instrumentation-psycopg2 = "^0.59b0"

Gunicorn Config

"""
The OpenTelemetry Python SDK uses the Global Interpreter Lock (GIL), which
can cause performance issues with Gunicorn spawn multiple processes
to serve requests in parallel.

To address this, register a post fork hook that runs after each worker
process is forked. Gunicorn post_fork hook initializes OTel inside each
worker which avoids lock inheritance/deadlocks and the perf hit.

Read more - https://grafana.com/docs/opentelemetry/instrument/python/?pg=blog&plcmt=body-txt#global-interpreter-lock
"""

import os
import multiprocessing
from uuid import uuid4

from opentelemetry import metrics, trace

from opentelemetry.exporter.otlp.proto.grpc.metric_exporter import (
    OTLPMetricExporter,
)
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import (
    OTLPSpanExporter,
)

# support for logs is currently experimental
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader
from opentelemetry.sdk.resources import Resource
from opentelemetry.sdk.resources import SERVICE_INSTANCE_ID
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor

from opentelemetry.instrumentation.django import DjangoInstrumentor
from opentelemetry.instrumentation.psycopg2 import Psycopg2Instrumentor

# your gunicorn config here
wsgi_app = "spectra.wsgi:application"
bind = "0.0.0.0:8000"
name = "spectra-backend"

# Multi-process concurrency (bypasses the GIL)
workers = multiprocessing.cpu_count()  # adjust via WEB_CONCURRENCY if needed
threads = 2
worker_class = "gthread"

# Performance & stability
timeout = 300
keepalive = 5
max_requests = 1000
max_requests_jitter = 100
worker_tmp_dir = "/dev/shm"

# Logging
loglevel = "debug"
accesslog = "-"
errorlog = "-"
capture_output = False

def post_fork(server, worker):
    server.log.info("Worker spawned (pid: %s)", worker.pid)

    collector_endpoint = os.environ.get("OTEL_EXPORTER_OTLP_ENDPOINT")

    resource = Resource.create(
        attributes={
            # each worker needs a unique service.instance.id to distinguish the created metrics
            SERVICE_INSTANCE_ID: str(uuid4()),
            "worker": worker.pid,
        }
    )

    tracer_provider = TracerProvider(resource=resource)
    tracer_provider.add_span_processor(
        BatchSpanProcessor(OTLPSpanExporter(endpoint=collector_endpoint))
    )
    trace.set_tracer_provider(tracer_provider)

    meter_provider = MeterProvider(
        resource=resource,
        metric_readers=[
            PeriodicExportingMetricReader(
                OTLPMetricExporter(endpoint=collector_endpoint)
            )
        ],
    )
    metrics.set_meter_provider(meter_provider)

    # Check lib.logs.OLTPLoggingHandler handler for logging instrumentation

    # Instruments database calls made through psycopg2
    Psycopg2Instrumentor().instrument(enable_commenter=True)

    # Instruments incoming HTTP requests handled by Django/Gunicorn
    DjangoInstrumentor().instrument(is_sql_commentor_enabled=True)


    server.log.info("OTel initialized in worker pid=%s", worker.pid)

Thank you!!