| === Wyly0 is now known as Wyly | ||
| manis05 | Hello, is there any logging that helps us know if we're hitting gunicorn worker limits inside docker? I read that [error] Worker timeout is usually an indicator of that. Aside from that, is there anything else? | 08:11 |
|---|---|---|
| patdk | hmm, that has nothing to do with docker | 08:29 |
| patdk | what exactly is a worker limit inside docker though | 08:30 |
| patdk | you start gunicorn with x workers, and that is how many it uses | 08:30 |
| patdk | if you made too many workers, or your cpu/disk/... isn't fast enough, you get a timeout | 08:31 |
| patdk | https://stackoverflow.com/questions/10855197/gunicorn-how-to-resolve-worker-timeout | 08:31 |
| manis05 | patdk thank you for responding. Yeah, we're running a python app through gunicorn inside docker. I think to state my question better - How can we determine if we have enough gunicorn workers to handle incoming requests? Is it logged somewhere? | 08:36 |
| patdk | no, and there is no way it would know that information | 08:37 |
| patdk | if your running a reverse proxy/loadbalancer infront of it (apache/nginx/...) | 08:37 |
| patdk | then you would look for errors from them, saying your backend (gunicorn) is not responding to requests | 08:38 |
| patdk | looking at 5xx errors | 08:38 |
| manis05 | patdk Nice, this is exactly what I'm looking for. Thanks again, I really appreciate it. | 08:40 |
| === ajfriesen164 is now known as ajfriesen16 | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!