Eliminate Warning Logs When Prometheus Metrics are Enabled

[haywoodsh]

Describe the bug

Enabling Prometheus metrics in the Ingress Controller generates unwarranted warning messages in the logs.

To Reproduce
Steps to reproduce the behavior:

1. Initiate Ingress Controller with the -enable-prometheus-metrics flag.
2. To expose the Prometheus metrics endpoint locally, execute:

kubectl port-forward `kubectl get po -n nginx-ingress -o=jsonpath="{.items[0].metadata.name}"` 9113:9113
--namespace=nginx-ingress

3. Navigate to http://localhost:9113 to access Prometheus metrics.
4. Observe that the Prometheus metrics display empty labels for resource_name, resource_namespace, and resource_type.
   Example:

lnginx_ingress_nginxplus_worker_connection_idle{class="nginx",id="0",pid="43",resource_name="",resource_namespace="",resource_type=""} 1

5. Warning messages similar to the following are logged~

level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=0
level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=1
level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=2
level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=3
level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=4
level=warn ts=2024-01-17T23:12:54.546Z caller=nginx_plus.go:1258 wrongnumberoflabelsforworker%v.Forlabels%v,gotvalues:%v.Emptylabelswillbeusedinstead=5

Expected behavior
Prometheus metrics should not contain empty labels, and no warning messages related to this should be logged.

Your environment

• NIC 3.4.1
• NGINX Plus

Additional context
Remove the variable label by modifying the line in the source code at https://github.com/nginxinc/kubernetes-ingress/blob/main/cmd/nginx-ingress/main.go#L703 from

workerPIDVariableLabels := []string{"resource_type", "resource_name", "resource_namespace"}`

to

var workerPIDVariableLabels []string

This change should result in metrics formatted as follows and eliminate the warning messages:

nginx_ingress_nginxplus_worker_connection_idle{class="nginx",id="0",pid="43"} 1

[github-actions]

Hi @haywoodsh thanks for reporting!

Be sure to check out the docs and the Contributing Guidelines while you wait for a human to take a look at this 🙂

Cheers!