celeryradar-sdk 0.2.2

Celery monitoring SDK — task events, worker heartbeats, beat schedules, and queue depths.

celeryradar-sdk

Celery monitoring SDK for CeleryRadar. Hooks Celery's standard signals to ship task events, worker heartbeats, beat schedules, and queue depth.

What gets monitored

• Task events — start, success, failure, retry, runtime, exception type
• Worker heartbeats — online/offline detection per worker hostname
• Beat schedules — fires, misses, and drift from the expected interval
• Queue depth — Redis broker depth per queue (single-instance only; RabbitMQ, SQS, and Redis Cluster planned)

Install

Requires Python 3.9+ and Celery 5.0+.

pip install celeryradar-sdk

Use

Sign up at celeryradar.com to get your API key, then:

import celeryradar_sdk

celeryradar_sdk.connect(api_key="cr_...", app_name="myapp")

That's it. The SDK is async and non-blocking — if the ingest endpoint is slow or unreachable, your workers don't notice; events drop with a warning rather than back up your task queue.

app_name is required. If you run multiple Celery apps under one API key — even on separate Redis brokers — give each a distinct app_name. It scopes the queue-depth poller's broker-side leader lock and disambiguates colliding queue names in the dashboard (two apps that both have a queue called celery stay separate by app).

Configuration

Common options:

celeryradar_sdk.connect(
    api_key="cr_...",
    app_name="myapp",
    capture_args=False,         # don't send task args/kwargs (default True)
    capture_exceptions=False,   # don't send exception/traceback text (default True)
    worker_name="api-worker-1", # override hostname; useful in k8s/Docker
    broker_url="redis://...",   # override app.conf.broker_url for the depth poller
)

CELERYRADAR_WORKER_NAME is also read from the environment and takes precedence over worker_name=. See the configuration docs for the full reference.

How it differs from Flower

Flower is a real-time inspector and admin tool — great for browsing the current task queue and revoking tasks. CeleryRadar is the persistence layer above it: history, alerts, dashboards, and trend analysis. They don't conflict.

Documentation

• Quick start — get connected in five minutes
• Configuration — every kwarg and env var
• How the SDK works — async ingest, retry queue, fork safety
• What gets monitored — signals and detection
• Troubleshooting — common issues and fixes

Background reading: Monitoring Celery in production walks through what to watch, what Celery gives you out of the box, and how the common options (Flower, Prometheus + Grafana, Sentry, Datadog) compare.

License

MIT