查询任务状态-Celery和Redis

Question

Okay so I have a relatively simple problem I think, and it's like I'm hitting a brick wall with it. I have a flask app, and a webpage that allows you to run a number of scripts on the server side using celery & redis(broker).

All I want to do, is when I start a task to give it a name/id (task will be portrayed as a button on the client side) ie

@app.route('/start_upgrade/<task_name>')
def start_upgrade(task_name):
    example_task.delay(1, 2, task_name=task_name)

Then after the task has kicked off I want to see if the task is running/waiting/finished in a seperate request, preferably like;

@app.route('/check_upgrade_status/<task_name>')
def get_task_status(task_name):
    task = celery.get_task_by_name(task_name)  
    task_state = task.state
    return task_state # pseudocode

But I can't find anything like that in the docs. I am very new to celery though just FYI so assume I know nothing. Also just to be extra obvious, I need to be able to query the task state from python, no CLI commands please.

Any alternative methods of achieving my goal of querying the queue are also welcome.

Answer 1

When you start a task with delay or apply_async an object AsyncResult is created and contains the id of the task. To get it you just have to store it in a variable.

For example

@app.route('/start_upgrade/<task_name>')
def start_upgrade(task_name):
    res = example_task.delay(1, 2, task_name=task_name)
    print res.id

You can store this id and maybe associate it with something else in a database (or just print it like I did in the example).

Then you can check the status of your task in a python console with :

from celery.result import AsyncResult
AsyncResult(your_task_id).status

Take a look at the result documentation you should get what you need there : http://docs.celeryproject.org/en/latest/reference/celery.result.html

Answer 2

I ended up figuring out a solution for my question from arthur's post.

In conjunction with redis I created these functions

import redis
from celery.result import AsyncResult

redis_cache = redis.StrictRedis(host='localhost', port=6379, db=0)

def check_task_status(task_name):
    task_id = redis_cache.get(task_name)
    return AsyncResult(task_id).status

def start_task(task, task_name, *args, **kwargs):
    response = task.delay(*args, **kwargs)
    redis_cache.set(task_name, response.id)

Which allows me to define specific names to tasks. Note I haven't actually tested this yet but it makes sense so.

Example usage;

start_task(example_task, "example_name", 1, 2)