Every other event pattern listening to the stream about django-eventstream HOT 10 OPEN

Hi,

Are you seeing this behavior when running in development mode with runserver? Or is this happening with a daphne deployment, and possibly you're running multiple instances?

from django-eventstream.

I'm running in development mode with runserver.

from django-eventstream.

I tried running the time example (in this repo) using the latest versions of all dependencies, including Django 3, and it seemed to work fine. Does it work for you too?

from django-eventstream.

I tried running the chat example, with Django 3 and it worked fine.

from django-eventstream.

Debugging further is difficult. The library could probably use some debug logging.

All I can suggest is to use curl to connect to the events endpoint in order to examine the raw stream. Also try to think about what is different with your app vs the chat example (use of threads, etc).

from django-eventstream.

I got the same result as before using curl.

I'm using a separate thread to call send_event(...). Do you think that can be the reason to observe this pattern? If yes, can you explain or do you know why that can be a problem?

from django-eventstream.

send_event is intended to be thread safe, and the time example calls it from a separate thread too, so it should be okay. But perhaps there is some extra complexity in your app that could be interfering. Are you sure send_event is actually getting called the number of times you expect? Are you using some kind of parallelization library that converts threads into processes?

from django-eventstream.

Can you trim your code down to a reproducible example and share it?

from django-eventstream.

I was sure that I was calling the send_event the number of times I was expecting.

I think I have the problem solved. But I'm not sure that I know exactly what was causing this issue.

The code that was calling the send_event was being initiated in the AppConfig.ready(). I don't know if this function was being initiated more than one time. Once I started calling python manage.py runserver with the option --noreload, the stream started receiving all the events as it should.

I think it is a solution for now, but I would like to understand better what was happening.
I would prefer to be able to use the runserver without the option --noreload.

Anyway, Thanks for your help!

from django-eventstream.

Glad to hear you've found a solution. I tried moving the time example's thread init to AppConfig.ready() and it still worked okay, even after modifying a file and letting it reload. An asgiref error was logged though.

Django's runserver reloading is tricky business and it wouldn't surprise me if there is some bug with it or in combination with Channels. Maybe upgrade your Daphne: django/channels#1374

from django-eventstream.