We are running an API server where users submit jobs for calculation, which take between 1 second and 1 hour. They then make requests to check the status and get their results, which could be (much) later, or even never.
Currently jobs are added to a pub/sub queue, and processed by various worker processes. These workers then send pub/sub messages back to a listener, which stores the status/results in a postgres database.
I am looking into using Celery to simplify things and allow for easier scaling.
Submitting jobs and getting results isn't a problem in Celery, using celery_app.send_task. However, I am not sure how to best ensure the results are stored when, particularly for long-running or possibly abandoned jobs.
Some solutions I considered include:
Give all workers access to the database and let them handle updates. The main limitation to this seems to be the db connection pool limit, as worker processes can scale to 50 replicas in some cases.
Listen to celery events in a separate pod, and write changes based on this to the jobs db. Only 1 connection needed, but as far as I understand, this would miss out on events while this pod is redeploying.
Only check job results when the user asks for them. It seems this could lead to lost results when the user takes too long, or slowly clog the results cache.
As in (3), but periodically check on all jobs not marked completed in the db. A tad complicated, but doable?
Is there a standard pattern for this, or am I trying to do something unusual with Celery? Any advice on how to tackle this is appreciated.
In the past I solved similar problem by modifying tasks to not only return result of the computation, but also store it into a cache server (Redis) right before it returns. I had a task that periodically (every 5min) collects these results and writes data (in bulk, so quite effective) to a relational database. It was quite effective until we started filling the cache with hundreds of thousands of results, so we implemented a tiny service that does this instead of task that runs periodically.
Related
I have tasks that do a get request to an API.
I have around 70 000 requests that I need to do, and I want to spread them out in 24 hours. So not all 70k requests are run at for example 10AM.
How would I do that in celery django? I have been searching for hours but cant find a good simple solution.
The database has a list of games that needs to be refreshed. Currently I have a cron that creates tasks every hour. But is it better to create a task for every game and make it repeat every hour?
The typical approach is to send them whenever you need some work done, no matter how many there are (even hundreds of thousands). The execution however is controlled by how many workers (and worker processes) you have subscribed to a dedicated queue. The key here is the dedicated queue - that is a common way of not allowing all workers start executing the newly created tasks. This goes beyond the basic Celery usage. You need to use celery multi for this use-case, or create two or more separate Celery workers manually with different queues.
If you do not want to over-complicate things you can use your current setup, but make these tasks with lowest priority, so if any new, more important, task gets created, it will be executed first. Problem with this approach is that only Redis and RabbitMQ backends support priorities as far as I know.
I'm doing some metric analysis on on my web app, which makes extensive use of celery. I have one metric which measures the full trip from a post_save signal through a celery task (which itself calls a number of different celery tasks) to the end of that task. I've been hitting the server with up to 100 requests in 5 seconds.
What I find interesting is that when I hit the server with hundreds of requests (which entails thousands of celery worker processes being queued), the time it takes for the trip from post save to the end of the main celery task increases significantly, even though I never do any additional database calls, and none of the celery tasks should be blocking the main task.
Could the fact that there are so many celery tasks in the queue when I make a bunch of requests really quickly be slowing down the logic in my post_save function and main celery task? That is, could the processing associated with getting the sub-tasks that the main celery task creates onto a crowded queue be having a significant impact on the time it takes to reach the end of the main celery task?
It's impossible to really answer your question without an in-depth analysis of your actual code AND benchmark protocol, and while having some working experience with Python, Django and Celery I wouldn't be able to do such an in-depth analysis. Now there are a couple very obvious points :
if your workers are running on the same computer as your Django instance, they will compete with Django process(es) for CPU, RAM and IO.
if the benchmark "client" is also running on the same computer then you have a "heisenbench" case - bombing a server with 100s of HTTP request per second also uses a serious amount of resources...
To make a long story short: concurrent / parallel programming won't give you more processing power, it will only allow you to (more or less) easily scale horizontally.
I'm not sure about slowing down, but it can cause your application to hang. I've had this problem where one application would backup several other queues with no workers. My application could then no longer queue messages.
If you open up a django shell and try to queue a task. Then hit ctrl+c. I can't quite remember what the stack trace should be, but if you post it here I could confirm it.
I have a "queue" of about a million entities on google app engine. I have to "pop" items off of the queue by using a query.
There are a bunch of client processes running all over the place that are constantly making requests to the stack. My problem is that when one of the clients requests an item, I want to make sure that I am removing that item from the front of the queue, sending it to that client process, and no other processes.
Currently, I am querying for the item, modifying its properties so that a query to the queue no longer includes that item, then saving the item. Using this method, it is very common for one item to be sent to more than one client process at the same time. I suspect this is because there is a delay to when I am making the writes and when they are being reflected to other processes.
Perhaps I need to be using transactions in some way, but when I looked into that, there were a couple of "gotchas". What is a good way to approach this problem?
Is there any reason not to implement the "queue" using App Engine's TaskQueue API? If size of the queue is the problem, TaskQueue could contain up to 200 million Tasks for a paid app, so a million entities would be easily handled.
If you want to be able to simulate queries for a certain task in the queue, you could use task tags, and have your client process pull tasks with a certain tag to be processed. Note that pulling tasks is supported through pull queues rather than push queues.
Other than that, if you want to keep your "queue-as-entities" implementation, you could use the Memcache API to signal the client process which entity need to be processed. Memcache provides stronger consistency when you need to share data between instances of your app compared to the eventual consistency of the HRD datastore, with the caveat that data in Memcache could be lost at any point in time.
I see two ways to tackle this:
What you are doing is ok, you just need to use transactions. If your processes are longer then 30s then you can offload them to task queue, which can be a part of transaction.
You could use Pull Queues, where you fill up a queue and than client processes pull tasks from the queue in atomic fashion (lease-delete cycle). With Pull Queues you can be sure that task is leased only once. Also task must be manually deleted from queue after it's done, meaning if your process dies task will be put back in queue after lease expires.
I have about 1000-10000 jobs which I need to run on a constant basis each minute or so. Sometimes new job comes in or other needs to be cancelled but it's rare event. Jobs are tagged and must be disturbed among workers each of them processes only jobs of specific kind.
For now I want to use cron and load whole database of jobs in some broker -- RabbitMQ or beanstalkd (haven't decided which one to use though).
But this approach seems ugly to me (using timer to simulate infinity, loading the whole database, etc) and has the disadvantage: for example if some kind of jobs are processed slower than added into the queue it may be overwhelmed and message broker will eat all ram, swap and then just halt.
Is there any other possibilities? Am I not using right patterns for a job? (May be I don't need queue or something..?)
p.s. I'm using python if this is important.
You create your initial batch of jobs and add them to the queue.
You have n-consumers of the queue each running the jobs. Adding consumers to the queue simply round-robins the distribution of jobs to each listening consumer, giving you arbitrary horizontal scalability.
Each job can, upon completion, be responsible for resubmitting itself back to the queue. This means that your job queue won't grow beyond the length that it was when you initialised it.
The master job can, if need be, spawn sub-jobs and add them to the queue.
For different types of jobs it is probably a good idea to use different queues. That way you can balance the load more effectively by having different quantities/horsepower of workers running the jobs from the different queues.
The fact that you are running Python isn't important here, it's the pattern, not the language that you need to nail first.
You can use asynchronous framework, e.g. Twisted
I don't think either it's a good idea to run script by cron daemon each minute (and you mentioned reasons), so I offer you Twisted. It doesn't give you benefit with scheduling, but you get flexibility in process management and memory sharing
I have my server on Google App Engine
One of my jobs is to match a huge set of records with another.
This takes very long, if i have to match 10000 records with 100.
Whats the best way of implementing this.
Im, using Web2py stack and deployed my application on Google App Engine.
maybe i'm misunderstanding something, but thos sounds like the perfect match for a task queue, and i can't see how multithreading will help, as i thought this only ment that you can serve many responses simultaneously, it won't help if your responses take longer than the 30 second limit.
With a task you can add it, then process until the time limit, then recreate another task with the remainder of the task if you haven't finished your job by the time limit.
Multithreading your code is not supported on GAE so you can not explicitly use it.
GAE itself can be multithreaded, which means that one frontend instance can handle multiple http requests simultaneously.
In your case, best way to achieve parallel task execution is Task Queue.
The basic structure for what you're doing is to have the cron job be responsible for dividing the work into smaller units, and executing each unit with the task queue. The payload for each task would be information that identifies the entities in the first set (such as a set of keys). Each task would perform whatever queries are necessary to join the entities in the first set with the entities in the second set, and store intermediate (or perhaps final) results. You can tweak the payload size and task queue rate until it performs the way you desire.
If the results of each task need to be aggregated, you can have each task record its completion and test for whether all tasks are complete, or just have another job that polls the completion records, to fire off the aggregation. When the MapReduce feature is more widely available, that will be a framework for performing this kind of work.
http://www.youtube.com/watch?v=EIxelKcyCC0
http://code.google.com/p/appengine-mapreduce/