I have an web app in which I am trying to use celery to load background tasks from a database. I am currently loading the database upon request, but would like to load the tasks on an hourly interval and have them work in the background. I am using flask and am coding in python.I have redis running as well.
So far using celery I have gotten the worker to process the task and the beat to send the tasks to the worker on an interval. But I want to retrieve the results[a dataframe or query] from the worker and if the result is not ready then it should load the previous result of the worker.
Any ideas on how to do this?
Edit
I am retrieving the results from a database using sqlalchemy and I am rendering the results in a webpage. I have my homepage which has all the various links which all lead to different graphs which I want to be loaded in the background so the user does not have to wait long loading times.
The Celery Task is being executed by a Worker, and it's Result is being stored in the Celery Backend.
If I get you correctly, then I think you got few options:
Ignore the result of the graph-loading-task, store what ever you need, as a side effect of the task, in your database. When needed, query for the most recent result in that database. If the DB is Redis, you may find ZADD and ZRANGE suitable. This way you'll get the new if available, or the previous if not.
You can look for the result of a task if you provide it's id. You can do this when you want to find out the status, something like (where celery is the Celery app): result = celery.AsyncResult(<the task id>)
Use callback to update farther when new result is ready.
Let a background thread wait for the AsyncResult, or native_join, which is supported with Redis, and update accordingly (not recommended)
I personally used option #1 in similar cases (using MongoDB) and found it to be very maintainable and flexible. But possibly, due the nature of your UI, option #3 will more suitable for you needs.
Related
I'm currently leveraging celery for periodic tasks. I am new to celery. I have two workers running two different queues. One for slow background jobs and one for jobs user's queue up in the application.
I am monitoring my tasks on datadog because it's an easy way to confirm my workers a running appropriately.
What I want to do is after each task completes, record which queue the task was completed on.
#after_task_publish.connect()
def on_task_publish(sender=None, headers=None, body=None, **kwargs):
statsd.increment("celery.on_task_publish.start.increment")
task = celery.tasks.get(sender)
queue_name = task.queue
statsd.increment("celery.on_task_publish.increment", tags=[f"{queue_name}:{task}"])
The following function is something that I implemented after researching the celery docs and some StackOverflow posts, but it's not working as intended. I get the first statsd increment but the remaining code does not execute.
I am wondering if there is a simpler way to inspect inside/after each task completes, what queue processed the task.
Since your question says is there a way to inspect inside/after each task completes - I'm assuming you haven't tried this celery-result-backend stuff. So you could check out this feature which is provided by Celery itself : Celery-Result-Backend / Task-result-Backend .
It is very useful for storing results of your celery tasks.
Read through this => https://docs.celeryproject.org/en/stable/userguide/configuration.html#task-result-backend-settings
Once you get an idea of how to setup this result-backend, Search for result_extended key (in the same link) to be able to add queue-names in your task return values.
Number of options are available - Like you can setup these results to go to any of these :
Sql-DB / NoSql-DB / S3 / Azure / Elasticsearch / etc
I have made use of this Result-Backend feature with Elasticsearch and this how my task results are stored :
It is just a matter of adding few configurations in settings.py file as per your requirements. Worked really well for my application. And I have a weekly cron that clears only successful results of tasks - since we don't need the results anymore - and I can see only failed results (like the one in image).
These were main keys for my requirement : task_track_started and task_acks_late along with result_backend
I need your opinion on a challenge that I'm facing. I'm building a website that uses Django as a backend, PostgreSQL as my DB, GraphQL as my API layer and React as my frontend framework. Website is hosted on Heroku. I wrote a python script that logs me in to my gmail account and parse few emails, based on pre-defined conditions, and store the parsed data into Google Sheet. Now, I want the script to be part of my website in which user will specify what exactly need to be parsed (i.e. filters) and then display the parsed data in a table to review accuracy of the parsing task.
The part that I need some help with is how to architect such workflow. Below are few ideas that I managed to come up with after some googling:
generate a graphQL mutation that stores a 'task' into a task model. Once a new task entry is stored, a Django Signal will trigger the script. Not sure yet if Signal can run custom python functions, but from what i read so far, it seems doable.
Use Celery to run this task asynchronously. But i'm not sure if asynchronous tasks is what i'm after here as I need this task to run immediately after the user trigger the feature from the frontend. But i'm might be wrong here. I'm also not sure if I need Redis to store the task details or I can do that on PostgreSQL.
What is the best practice in implementing this feature? The task can be anything, not necessarily parsing emails; it can also be importing data from excel. Any task that is user generated rather than scheduled or repeated task.
I'm sorry in advance if this question seems trivial to some of you. I'm not a professional developer and the above project is a way for me to sharpen my technical skills and learn new techniques.
Looking forward to learn from your experiences.
You can dissect your problem into the following steps:
User specifies task parameters
System executes task
System displays result to the User
You can either do all of these:
Sequentially and synchronously in one swoop; or
Step by step asynchronously.
Synchronously
You can run your script when generating a response, but it will come with the following downsides:
The process in the server processing your request will block until the script is finished. This may or may not affect the processing of other requests by that same server (this will depend on the number of simultaneous requests being processed, workload of the script, etc.)
The client (e.g. your browser) and even the server might time out if the script takes too long. You can fix this to some extent by configuring your server appropriately.
The beauty of this approach however is it's simplicity. For you to do this, you can just pass the parameters through the request, server parses and does the script, then returns you the result.
No setting up of a message queue, task scheduler, or whatever needed.
Asynchronously
Ideally though, for long-running tasks, it is best to have this executed outside of the usual request-response loop for the following advantages:
The server responding to the requests can actually serve other requests.
Some scripts can take a while, some you don't even know if it's going to finish
Script is no longer dependent on the reliability of the network (imagine running an expensive task, then your internet connection skips or is just plain intermittent; you won't be able to do anything)
The downside of this is now you have to set more things up, which increases the project's complexity and points of failure.
Producer-Consumer
Whatever you choose, it's usually best to follow the producer-consumer pattern:
Producer creates tasks and puts them in a queue
Consumer takes a task from the queue and executes it
The producer is basically you, the user. You specify the task and the parameters involved in that task.
This queue could be any datastore: in-memory datastore like Redis; a messaging queue like RabbitMQ; or an relational database management system like PostgreSQL.
The consumer is your script executing these tasks. There are multiple ways of running the consumer/script: via Celery like you mentioned which runs multiple workers to execute the tasks passed through the queue; via a simple time-based job scheduler like crontab; or even you manually triggering the script
The question is actually not trivial, as the solution depends on what task you are actually trying to do. It is best to evaluate the constraints, parameters, and actual tasks to decide which approach you will choose.
But just to give you a more relevant guideline:
Just keep it simple, unless you have a compelling reason to do so (e.g. server is being bogged down, or internet connection is not reliable in practice), there's really no reason to be fancy.
The more blocking the task is, or the longer the task takes or the more dependent it is to third party APIs via the network, the more it makes sense to push this to a background process add reliability and resiliency.
In your email import script, I'll most likely push that to the background:
Have a page where you can add a task to the database
In the task details page, display the task details, and the result below if it exists or "Processing..." otherwise
Have a script that executes tasks (import emails from gmail given the task parameters) and save the results to the database
Schedule this script to run every few minutes via crontab
Yes the above has side effects, like crontab running the script in multiple times at the same time and such, but I won't go into detail without knowing more about the specifics of the task.
I have a Django application that uses large data structures in-memory (due to performance constraints). This wouldn't be a problem, but I'm using Heroku, where if the python web process takes more than 30s to start, it will be stopped as it's considered a timeout error. Because of the problem aforementioned, I've used a daemon process(worker in Heroku) to handle the construction of the data structures and Redis to handle the message passing between processes.
When the worker finishes(approx 1 minute), it stores the data structures(50Mb or so) in Redis.
And now comes the crux of the matter...Django follows the request/response paradigm and it's synchronised. This implies a Django view should exist to handle the callback from the worker announcing it's done. Even if I use something fancier like a pub/sub from Redis, I'm still forced to evaluate the queue populated by a publisher in a view.
How can I circumvent the necessity of using a Django view? Isn't there an async way of doing this?
Below is the solution where I use a pub/sub inside a view. This seems bad, but I can't think of another way.
views.py
...
# data_handler can enqueue tasks on the default queue
data_handler = DataHandler()
strict_redis = redis.from_url(settings.DEFAULT_QUEUE)
pub_sub = strict_redis.pubsub()
# this puts the job of constructing the large data structures
# on the default queue so a worker can pick it up. Being async,
# it returns with an empty set of data structures.
data_structures = data_handler.start()
pub_sub.subscribe(settings.FINISHED_DATA_STRUCTURES_CHANNEL)
#require_http_methods(['POST'])
def store_and_fetch(request):
user_data = json.load(request.body.decode('utf8'))
message = pub_sub.get_message()
if message:
command = message['data'] if 'data' in message else ''
if command == settings.FINISHED_DATA_STRUCTURES_INIT.encode('utf-8'):
# this takes the data from redis and updates data_structures
data_handler.update(data_structures)
return HttpResponse(compute_response(user_data, data_structures))
Update: After working for multiple months with this, I can now say it's definitely better(and wiser) NOT to fiddle with Django's request/response cycle. There are things like Django RQ Scheduler, or Celery that can do async tasks just fine. If you want to update the main web process after some repeatable job completed, it's simpler to use something like python requests package, sending a POST to the web process from the worker that did the scheduled job. In this way we don't circumvent Django's mechanisms, and more importantly, it's simpler to do overall.
Regarding the Heroku constraints I mentioned at the beginning of the post. At the moment I wrote this question I was quite a newbie with heroku and didn't know much about the release phase. In the release phase we can set up all the complex logic we need for the main process. Thus, at the end of the release phase, we simply need to notify the web process, in the manner I've described above and use some distributed memory buffer (even Redis will work just fine).
Is there any common solution to store and reuse celery task results without executing tasks again? I have many http fetch tasks in my metasearch project and wish to reduce number of useless http requests (they can take long time and return same results) by store results of first one and fire it back without real fetching. Also it will be very useful to does not start new fetch task when the same one is already in progress. Instead of running new job app has to return AsyncResult by id (id is unique and generated by task call args) of already pending task.
Looks like I need to define new apply_async(Celery.send_task) behavior for tasks with same task_id:
if task with given task_id doesn't started yet then start it
if task with given task_id already started return AsyncResult(task_id) without actually run task
#task decorator should accept new ttl
kwarg to determine cache time (only for redis backend?)
Looks like the simplest answer is to store your results in a cache (like a database) and first ask for the result from your cache else fire the http request.
I don't think there's something specific to celery that can perform this.
Edit:
To comply with the fact that you the tasks are sent at the same time an additional thing would be to build a lock for celery task (see Celery Task Lock receipt).
In your case you want to give the lock a name containing the task name and the url name. And you can use whatever system you want for cache if visible by all your workers (Redis in your case?)
I'm wondering what kind of options there are for monitoring celery tasks from a browser, after they have been deployed to a worker?
My current application stack is a flask app running inside twisted, using celery to run dozens to thousands of small background tasks (updating metadata in a repository, creating image derivatives, etc.) I'm envisioning using ajax long-polling to monitor the status of the celery tasks initiated by the user. I'm using redis for the backend broker and results.
I see celery has some command line ways to monitor tasks, or flower for a web dashboard. But if I wanted to see more detailed status from a particular task sent to celery, would it make more sense for that task to print / write to a log file, then long-poll that file for changes from the flask front-end?
At this point a user can say, "update these 10,000 items", the tasks are sent to celery, and the front-end very quickly says, "job sent!". And the tasks do complete. But I'd like to have the user navigate to "/status" and see the status of those 10,000 small jobs - even a scrolling log file would probably work.
Any suggestions would be greatly appreciated. Took a lot of head scratching to make it this far sketching things out, but I'm spinning my wheels figuring out exactly WHAT to long-poll from the user front-end.
Try Jobstatic, which is extending Celery.
From project description:
Jobtastic gives you goodies like:
Easy progress estimation/reporting
Job status feedback
Helper methods for gracefully handling a dead task broker (delay_or_eager and delay_or_fail)
Super-easy result caching
Thundering herd avoidance
Integration with a celery jQuery plugin for easy client-side progress display
Memory leak detection in a task run
Jobtastic was a great idea, but not quite what worked for us. In the end, decided to create an incrementing job number (stored in Redis alongside results and broker), push all celery task id's associated with that job number into a python object, then pickle and store that in redis. We can then use that later to see if the entire "job" is complete, or the status thereof. For our purposes, works just lovely.