Django based Docker image failing to start

1

Im trying to build a one click installer for our platform, Flagsmith. This image works fine whenever we have run it in other environments: locally, Fargate, Digital Ocean apps etc.

One question I couldn’t get an answer to - do you need to specify the port that the docker image exposes or is this automatically handled?

On Render I’m seeing this in our logs:

Jan 11 09:59:33 AM  [2022-01-11 09:59:33 +0000] [74] [INFO] Worker exiting (pid: 74)
Jan 11 09:59:33 AM  [2022-01-11 09:59:33 +0000] [75] [INFO] Worker exiting (pid: 75)
Jan 11 09:59:33 AM  [2022-01-11 09:59:33 +0000] [76] [INFO] Worker exiting (pid: 76)
Jan 11 09:59:34 AM  [2022-01-11 09:59:34 +0000] [77] [INFO] Booting worker with pid: 77
Jan 11 09:59:34 AM  [2022-01-11 09:59:34 +0000] [78] [INFO] Booting worker with pid: 78
Jan 11 09:59:34 AM  [2022-01-11 09:59:34 +0000] [79] [INFO] Booting worker with pid: 79
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:97)
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:98)
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:99)
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [97] [INFO] Worker exiting (pid: 97)
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [98] [INFO] Worker exiting (pid: 98)
Jan 11 09:58:58 AM  [2022-01-11 09:58:58 +0000] [99] [INFO] Worker exiting (pid: 99)
Jan 11 09:58:59 AM  [2022-01-11 09:58:59 +0000] [100] [INFO] Booting worker with pid: 100
Jan 11 09:58:59 AM  [2022-01-11 09:58:59 +0000] [101] [INFO] Booting worker with pid: 101
Jan 11 09:58:59 AM  [2022-01-11 09:58:59 +0000] [102] [INFO] Booting worker with pid: 102
Jan 11 09:59:29 AM  [2022-01-11 09:59:29 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:100)
Jan 11 09:59:29 AM  [2022-01-11 09:59:29 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:101)
Jan 11 09:59:29 AM  [2022-01-11 09:59:29 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:102)

Any idea what might be causing this?

Our render.yaml is here.

2

Hello Ben,

Port detection should be handled automatically for your service - you can check that the correct port was detected by looking at the internal address section near the top of your service’s page.

As for the logs you’re seeing, I can see that you added an increased timeout for Gunicorn - did things improve after adding that timeout?