[BUG] Gunicorn workers out of memory in prod mode

[borolepratik]

Describe the bug
I have a long running task that runs for about ~250 seconds and works as expected in dev mode. In prod mode, it seems to run out of memory?

[2025-01-05 22:33:57 +0530] [12395] [CRITICAL] WORKER TIMEOUT (pid:12787)
[2025-01-05 22:33:58 +0530] [12395] [ERROR] Worker (pid:12787) was sent SIGKILL! Perhaps out of memory?

Specifics (please complete the following information):

• Python Version: 3.13.1
• Reflex Version: 0.6.7
• OS: Windows 11 (WSL), Macos 15.2

[linear]

ENG-4300 [BUG] Gunicorn workers out of memory in prod mode

[borolepratik]

there yet may be more issues to be iron out for the prod mode. eg #3605

[adhami3310]

hey there, can you test it with granian? you can do so by setting the environment variable REFLEX_USE_GRANIAN=1. If there's minimal repro code that could be great.

[borolepratik]

using REFLEX_USE_GRANIAN=1 has solved the problem for me. thank you! tested on Windows WSL and reflex v0.6.8.
apologies for no repro code.