Daphne ASGI Django Long running requests cause upstream connection refused NGINX

64 views

Skip to first unread message

Bernsy

unread,

Sep 9, 2021, 1:01:02 PM9/9/21

to Django users

I have a Django 3.2.4 application which I am running on Kubernetes, fronted by NGINX, served via ASGI / daphne. I have a long running API call (no longer than 30 seconds) that when running, seems to cause other requests to be refused.

In the NGINX logs I see connect() failed (111: Connection refused) while connecting to upstream and in the pods I see the liveness probe fails (because connection refused), and get Killed 8 pending application instances.