Connect Celery to Bugsink
When a Celery task fails, Bugsink shows you the exception, the application frame that raised it, and the name of the task the worker was running.
This guide covers the initial setup and Celery’s automatic capture of failed tasks. You’ll configure the SDK where your worker loads its Celery app, then queue a deliberate failure through a real worker to confirm it reaches Bugsink.
To follow along, create a Bugsink project and copy its DSN. The quickstart guide shows where to find it.
Install the SDK
Install the open-source Sentry SDK with its Celery support in the environment your worker uses. Bugsink accepts the Sentry event protocol, so the SDK can send task failures directly to your Bugsink project.
python -m pip install --upgrade "sentry-sdk[celery]"python -m pip install --upgrade "sentry-sdk[celery]"
Configure the Celery worker
Initialize the SDK in the module that creates your Celery app. The worker must import this module during startup, before it begins processing tasks:
import os import sentry_sdk from celery import Celery from sentry_sdk.integrations.celery import CeleryIntegrationimport os import sentry_sdk from celery import Celery from sentry_sdk.integrations.celery import CeleryIntegration
Read the DSN from the environment and enable the Celery integration before creating the app:
sentry_sdk.init( dsn=os.environ["SENTRY_DSN"], integrations=[CeleryIntegration()], )sentry_sdk.init( dsn=os.environ["SENTRY_DSN"], integrations=[CeleryIntegration()], )
Set SENTRY_DSN in the environment where the worker runs. Use the DSN copied
from your Bugsink project:
export SENTRY_DSN=https://key@bugsink.example/1export SENTRY_DSN=https://key@bugsink.example/1
Queue a task that fails
A task does not need an SDK call. If it raises an exception and Celery marks
the result as failed, the integration reports that exception from the worker.
This inventory task raises ProductNotFound for an unknown product:
@app.task(name="inventory.reserve_product") def reserve_product(sku, quantity=1): try: product = INVENTORY[sku] except KeyError: raise ProductNotFound(f"Product {sku} is not in inventory") from None if product["stock"] < quantity: raise ValueError(f"Only {product['stock']} units of {sku} are available") return { "sku": sku, "name": product["name"], "reserved": quantity, }@app.task(name="inventory.reserve_product") def reserve_product(sku, quantity=1): try: product = INVENTORY[sku] except KeyError: raise ProductNotFound(f"Product {sku} is not in inventory") from None if product["stock"] < quantity: raise ValueError(f"Only {product['stock']} units of {sku} are available") return { "sku": sku, "name": product["name"], "reserved": quantity, }
Start your worker in one terminal as usual:
celery -A inventory_worker.celery_app worker --loglevel=infocelery -A inventory_worker.celery_app worker --loglevel=info
In another process, queue the task rather than calling its Python function directly:
from inventory_worker.tasks import reserve_product result = reserve_product.delay("celery-missing") print(result.get(timeout=30, propagate=False))from inventory_worker.tasks import reserve_product result = reserve_product.delay("celery-missing") print(result.get(timeout=30, propagate=False))
Celery records a FAILURE result and the worker logs the unhandled task
exception.
See the failed task in Bugsink
Open your Bugsink project and select the new ProductNotFound issue. Its event
points to the task body and uses inventory.reserve_product as the task and
transaction name.
ProductNotFound
That’s it: failed Celery tasks will now be reported to Bugsink.
The integration runs inside the worker, so producers do not need to catch task
errors or wait for results to report them. Each event includes the exception,
the worker-side Python stack, the task name, and a celery_task_id tag that you
can use to relate it to Celery’s result or logs.
Limitations
With the SDK’s default data settings, Celery task arguments and keyword
arguments appear as [Filtered]. The task name, task ID, exception, and stack
remain available without sending the values passed to the task.
For handled exceptions, logging, tags, structured context, and other SDK options, see the broader Python guide.
