07-Jul-2022 02:28
Hi,
We find in two kubernetes clusters that any ingress nginx config changes are made (new backend, log modify...), the result is that the the proxy response time will be bad, slow, as long as we do not delete the pods. After restarting the ingress-nginx-controller pods everything is ok.
ingress-nginx-4.0.17
k8s: v1.21.6+2.el7
Idea?
07-Jul-2022 10:01
There are a lot of of articles about tracing the traffic in nginx:
https://www.nginx.com/blog/application-tracing-nginx-plus/
https://www.nginx.com/blog/opentracing-nginx-plus/
https://docs.nginx.com/nginx-ingress-controller/troubleshooting/troubleshoot-ingress-controller/
Also you can check the release notes for known issues and fixes and better run the latest nginx version:
https://docs.nginx.com/nginx-ingress-controller/releases/#nginx-ingress-controller-222
22-Jul-2022 14:41
Probably a silly question, but have you checked the logs of the ingress pods?
Also, have you checked any performance graphs in Prometheus to check for ie. CPU throttling?
13-Oct-2022 13:43
If you managed to get the needed answers, please flag the question as answered.