celeryviz 0.0.1

A tool for visualising execution of Celery tasks.

CeleryViz

A tool for visualising execution of Celery tasks.

Live Demo

This project simplifies debugging of asynchronous Celery tasks by offering a visual representation of the task execution flow. Instead of combing through the log files, developers can use Celeryviz to visually trace and debug task processes more efficiently.

Installation

pip install celeryviz

Usage

1. Create a celery project.

• Use this gist for a minimal example.

curl https://gist.githubusercontent.com/bhavya-tech/d937ef45905720014ee12fe332352966/raw/0afac784adfb6b407fa83ce4b19e6f3cab4d80d9/example_app.py -o example_app.py

2. Start the celery worker:

• Ensure a message broker is running (can use RabbitMQ for simplicity)

celery -A example_app worker

3. Start the CeleryViz server:

celery -A example_app celeryviz

• Open your browser and go to http://0.0.0.0:9095/app/

• The attach_log_sender in the example gist sends logs to CeleryViz.

  • If a custom logger does not support handlers, then monkeypatch the logger to mimic the attach_log_sender function.

Contributing

Setting up the development environment

1. Clone this repository.

git clone https://github.com/bhavya-tech/celeryviz.git

2. Build the webapp

• The UI webapp is maintained separately in celeryviz_with_lib.

• Run the following command to build the latest version of standard webapp locally (ensure that docker is installed):

docker build --output ./celeryviz/static ./build_ui

For customised builds, the following optional build args can be used:

• GITHUB_PAT
  • If any of dependency repo is private, add a github personal access token as a build argument.
• GIT_REPO
  • The URL of the repository to build.
  • Defaults to bhavya-tech/celeryviz_with_lib
• SOURCE
  • The branch of the repository to build. Default is main.
  • A particular commit hash can also be used.

docker build --output ./celeryviz/static --build-arg="GITHUB_PAT=<your github personal access token>" ./build_ui

3. Install the package in editable mode:

pip install -e .