Getting nginx_ingress_controller_requests metric

[JanCizmar]

Hi,

the kubernetes.io version of nginx ingress controller exposes the nginx_ingress_controller_requests metrics, which provides a status code via label.

Can I configure nginxinc controller to provide the same metric? I am getting random 502s and I need to monitor it.

Thanks!

[haywoodsh]

Hi @JanCizmar, the NGINX Ingress Controller exposes all the metrics from NGINX/NGINX Plus itself. You can see the list of exported metrics here. Could nginx_http_requests_total or nginxplus_http_requests_total be what you are looking for?

[JanCizmar]

Well, all the metrics labeled by status are part of NGINX plus, which I am currently not using.

[brianehlert]

That could be possible using access logging as described here: https://www.nginx.com/blog/using-nginx-logging-for-application-performance-monitoring/

[JanCizmar]

Thanks. I actually just used Loki to store the logs from nginx ingress controller and then used this query to get the rate of 502 responses.

sum(rate({container="nginx-ingress"} |~ "\"\\s502\\s"[$__interval]))

[brianehlert]

quick question -
do you scrape the access logs from the ingress controller pod logs?
Or do you use the NGINX native access log syslog streaming to send logs to a Loki collector?
Or, would you have a preference if you could?

[JanCizmar]

I am using controller pod logs. It was there out of the box.