# One request at a time from user

**URL:** <https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529>\
**Category:** Forms & APIs\
**Created:** [February 28, 2024, 10:01pm UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529 "2024-02-28T22:01:08Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![enzoferrari1](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/enzoferrari1/32/19037_2.png) [@enzoferrari1](https://forum.djangoproject.com/u/enzoferrari1)\
**Post date:** [February 28, 2024, 10:01pm UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529/1 "2024-02-28T22:01:08Z")

</div>

Hi! I’m making a project in Django. I would like to lock the user to make simultaneous requests. I’m making calls to other APIs and I’m afraid that if the user requests are made simultaneously then it could make more API calls than needed, costing me more. What’s the best tool for this? My first idea was defining a field in the user table that sets a flag that represents when the server is busy processing a user request, but I don’t think this is the best idea. I think this could be made from the middleware but I’m lost there. Should I use the server cache? Should I use postgresql locks? I don’t know what to do. Greetings.

---

<div class="post-metadata">

**Author:** ![KenWhitesell](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/kenwhitesell/32/280_2.png) [@KenWhitesell](https://forum.djangoproject.com/u/KenWhitesell)\
**Post date:** [February 28, 2024, 11:05pm UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529/2 "2024-02-28T23:05:17Z")

</div>

> [@enzoferrari1](#):
>
> Should I use the server cache?

Assuming all Django instances are sharing the same cache, this may be your best bet.

This has the advantage that you can set an expiration on the cache key, so that if the cache isn’t released by the other process for some reason, it will still expire in a reasonable time.

---

<div class="post-metadata">

**Author:** ![enzoferrari1](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/enzoferrari1/32/19037_2.png) [@enzoferrari1](https://forum.djangoproject.com/u/enzoferrari1)\
**Post date:** [February 29, 2024, 12:22am UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529/3 "2024-02-29T00:22:17Z")

</div>

Thanks, I’ll try and post the cache approach if I’m successful. Sorry if this sounds too dumb but, is it a bad idea to define a boolean flag in the user model to lock when he’s making a request (either modifying directly or defining a middleware) and unlocking it when the request finishes? From what I’ve understood, the cache is also stored in the database. What’s the difference between simply defining a flag and using the cache for this problem?  
Also, you mentioned ‘Assuming all Django instances are sharing the same cache’. Would that assumption be violated if I’m running the application on a cloud load balancer?

---

<div class="post-metadata">

**Author:** ![KenWhitesell](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/kenwhitesell/32/280_2.png) [@KenWhitesell](https://forum.djangoproject.com/u/KenWhitesell)\
**Post date:** [February 29, 2024, 12:47am UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529/4 "2024-02-29T00:47:29Z")

</div>

> [@enzoferrari1](#):
>
> From what I’ve understood, the cache is also stored in the database.

It _can_ be, but _shouldn’t_ be. See your options at [Django’s cache framework | Django documentation | Django](https://docs.djangoproject.com/en/5.0/topics/cache/#setting-up-the-cache)

> [@enzoferrari1](#):
>
> Would that assumption be violated if I’m running the application on a cloud load balancer?

Possibly. It depends upon how the cache is configured.

---

<div class="post-metadata">

**Author:** ![enzoferrari1](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/enzoferrari1/32/19037_2.png) [@enzoferrari1](https://forum.djangoproject.com/u/enzoferrari1)\
**Post date:** [February 29, 2024, 4:48am UTC](https://forum.djangoproject.com/t/one-request-at-a-time-from-user/28529/5 "2024-02-29T04:48:20Z")

</div>

Here’s the code that i made for the middleware in order to lock the user:

```auto
class RequestFlagMiddleware:
    def __init__ (self, get_response):
        self.get_response = get_response

    def __call__ (self, request):
        user_id = request.user.id if request.user.is_authenticated else None
        
        # Check if the user is currently making a request
        if user_id and cache.get(f'user_request_{user_id}_in_progress'):
            return JsonResponse({'message': 'Another request is already in progress'}, status=409) # Conflict

        # Set the flag indicating that the user is making a request
        if user_id:
            cache.set(f'user_request_{user_id}_in_progress', True, timeout = 300)

        response = self.get_response(request)

        # Clear the flag after processing the request
        if user_id:
            cache.delete(f'user_request_{user_id}_in_progress')

        return response

```

This is made for sync views. I’m in the process of understanding how the middleware works with async. Thank you very much Ken for the effort put in this forum, it’s appreciated.  
Greetings.