OpenTelemetry Celery Instrumentation

Instrument celery to trace Celery applications.

Usage

• Start broker backend

docker run -p 5672:5672 rabbitmq

• Run instrumented task

from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchExportSpanProcessor
from opentelemetry.instrumentation.celery import CeleryInstrumentor

from celery import Celery
from celery.signals import worker_process_init

@worker_process_init.connect(weak=False)
def init_celery_tracing(*args, **kwargs):
    trace.set_tracer_provider(TracerProvider())
    span_processor = BatchExportSpanProcessor(ConsoleSpanExporter())
    trace.get_tracer_provider().add_span_processor(span_processor)
    CeleryInstrumentor().instrument()

app = Celery("tasks", broker="amqp://localhost")

@app.task
def add(x, y):
    return x + y

add.delay(42, 50)

API

class opentelemetry.instrumentation.celery.CarrierGetter

Bases: opentelemetry.trace.propagation.textmap.DictGetter

get(carrier, key)

Function that can retrieve zero or more values from the carrier. In the case that the value does not exist, returns an empty list.

Parameters

• carrier – An object which contains values that are used to construct a Context.

• key – key of a field in carrier.

Returns: first value of the propagation key or an empty list if the

key doesn’t exist.

keys(carrier)

Function that can retrieve all the keys in a carrier object.

Parameters

carrier – An object which contains values that are used to construct a Context.

Returns

list of keys from the carrier.

class opentelemetry.instrumentation.celery.CeleryInstrumentor

Bases: opentelemetry.instrumentation.instrumentor.BaseInstrumentor