elery is digestive

@stalwartcoder

Agenda

Abhishek

πŸ‘¨β€πŸ’» Software Eng. at Essentia SoftServ

🐍 Pythonista

πŸ‘¨β€πŸ‘©β€πŸ‘§β€πŸ‘¦Β  Community first person πŸ’›

connect with me:

🚫 not "10x Engineer"

Task Queue

Task Queue

Manage background work - outside the usual HTTP request-response cycle.

For example, a web app poll the API every 10 minutes to collect the names of the top crypto currencies.

A task queue would handle invoking code to call the API, process the results and store them in a persistent DB for

later use.

Input to task queue's are a unit of workΒ  - separate code.

πŸ“¬

Task Queue is a system for parallel execution of tasks.

Need of Task Queue... πŸ•΅οΈ

✊ Useful in certain situation.

πŸ‘‰Β General guidelines :

  • Are you using a lot of cron jobs to process data in the background?
  • Does your application takes more than a few seconds to generate response?
  • Do you wish you could distribute the processing of the data generated by your application among many servers?
  • OR an asynchrounous DB update ?
distributes task

Broker

sends task

Client

Worker

Worker

Worker

distributes task
distributes task

WAIT!!

I’ve heard Asynchronous before!

Yes, AJAX

Some of the response time issues can be solved :

  • With AJAX responses that continually enhance the initial response.
  • Only if the AJAX responses also complete within a reasonable amount of time.

You need a Queue :

  • Long processing times can’t be avoided in generating responses.
  • You want application data to be continuously processed in the background and readily available when requested.

Different Frameworks or Projects for TQΒ 

πŸ”· RQ (Redis Queue)

πŸ”· Taskmaster

πŸ”· Huey

πŸ”· AWS SQS

πŸ”· CloudAMQP

Blah....blahh...blaahhh

Let's talk about *celery*

Do it....

Celery is an asynchronous task queue/job queue based on distributed message passing.

It is focused on real-time operation, and scheduling as well.

Tasks can execute asynchronously (in the background) or synchronously (wait until ready).

Easy to integrate & multi broker support .

Key concept

  • Message Broker
  • Worker - Executors

De facto standard for Python

Why should i use it ?

πŸ€”

it's just simple!
from celery import Celery

app = Celery('tasks', broker='pyamqp://guest@localhost//')

@app.task
def add(x, y):
    return x + y
>>> from tasks import add
>>> add.delay(4, 4)
<AsyncResult: 889143a6-39a2-4e52-837b-d80d33efb22d>

Define task

Execute task

$ celery -A tasks worker --loglevel=info

Run the celery worker

$ pip install celery

install celery

  • Minimize request/response cycle
  • Smoother user experience
  • Offload time/cpu intensive processes
  • Flexibility - many points of customization
  • Supports web hook, job routing
  • Remote control of workers
  • Logging, debugging & monitoring (Flower) support.
  • Wide support of web frameworks actively developed.
  • Amazing documentation and lot of tutorials!

πŸ› οΈ Framework Support

client 1

CeleryWorker 1

client 2

Broker

Task Result Storage

Task Queue 1
Task Queue 2
Task Queue N

...

send tasks

send tasks

distribute tasks

distribute tasks

get task result

store task result

store task result

Celery Architecture

CeleryWorker 2

Celery

Broker

DB for task result

RabbitMQ

Redis

AWS SQS

Selecting BrokerΒ 

Finally !! πŸ€¦β€β™‚οΈ

Approach for using celery πŸ‘¨β€πŸ’»

  • Simple setup : single queue & single worker​​
  • Large setup : multiple queue & dedicated worker with concurrency pattern​
  • Key choicesΒ :​​
  • Broker
  • Result store

More detail :

Some Tips 😬

  • Start simple !
  • Watch for the result .
  • Monitor it !Β 
  • Consume faster than you produce.
  • Try concurrency over tasks.
  • Keep your task clean.
  • Replace your cron jobs!

More:

Resources πŸ“š

Conclusion πŸ’‘

  • Take a slow task / function
  • Decouple it from system
  • Run it asynchronously
  • Don't use it for everything, they are not magic wand !
Made with Slides.com