mher/flower

Flower is a web-based dashboard for monitoring and managing Celery, a Python system that runs background tasks across one or more worker processes. When you have an application that offloads time-consuming work, like sending emails, processing images, or running reports, Celery handles the scheduling and Flower gives you a window into what is actually happening. The main interface shows live task activity pulled directly from Celery events. You can see which tasks are running, how long they have been running, their input arguments, start times, and outcomes. Task history is also available for reviewing completed or failed work. Beyond viewing, Flower lets you control your workers from the browser. You can shut down or restart worker instances, change the size of the worker pool, adjust which task queues a worker is consuming from, and apply time limits or rate limits to tasks. Individual tasks can be cancelled or forcibly stopped. Flower also monitors the message broker layer, which is typically Redis or RabbitMQ. This gives you statistics on queue lengths and message throughput. Authentication is supported through HTTP Basic Auth or via OAuth with Google, GitHub, GitLab, or Okta. A Prometheus integration lets you pull metrics into external monitoring systems if you use that. There is also a REST API that lets you perform control operations programmatically, without going through the browser. For example, you can restart a worker pool or trigger a task with a single HTTP request. Flower is installable via pip and also available as a Docker image. Documentation is hosted on Read the Docs. The project is licensed under the BSD 3-Clause License.