Since the problem described in this issue should be resolved in a recent advisory, it has been closed.

For information on the advisory (Important: OpenShift Container Platform 4.18.1 bug fix and security update), and where to find the updated files, follow the link below.

If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHSA-2024:6122

Marked it verified with 4.18.0-0.nightly-2024-11-22-231049, thanks.

verified it with the pre-merged test, please refer to https://github.com/openshift/route-controller-manager/pull/49#issuecomment-2431568891 for more detail, thanks.

The issue could be reproduced with 4.18.0-0.nightly-2024-09-21-014704

1.
% oc get clusterversion                                           
NAME      VERSION                              AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.18.0-0.nightly-2024-09-21-014704   True        False         6h45m   Cluster version is 4.18.0-0.nightly-2024-09-21-014704

2. create pod/service on ns test
% oc -n test get all
Warning: apps.openshift.io/v1 DeploymentConfig is deprecated in v4.14+, unavailable in v4.10000+
NAME                READY   STATUS    RESTARTS   AGE
pod/appach-server   1/1     Running   0          5h1m


NAME                  TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)     AGE
service/sec-apach2    ClusterIP   172.30.246.65   <none>        27443/TCP   5h1m
service/unsec-apach   ClusterIP   172.30.157.72   <none>        28080/TCP   5h1m

3.  create ingress ingress-reen1, ingress-pass1 and ingress-with-class, then delete ingress ingress-reen1
% oc -n test get ingress
NAME                 CLASS    HOSTS                                                   ADDRESS   PORTS     AGE
ingress-pass1        <none>   reen1-test.sec.shudi-8a23.qe.devcluster.openshift.com             80, 443   4h54m
ingress-with-class   mytest   foo.bar.com                                                       80        4h53m

4. navigate the web console Observe >> Metrics, input openshift_ingress_to_route_controller_ingress_without_class_name
the value for "ingress-reen1" was still 1

hi team any update on this issue?

I have the same Issue on a 4.14.3 cluster:

```

oc get ingress -A
No resources found

```

It's a red hat internal cluster I can provide access or an must-gather.

We also observed that the alerts are not differentiating between different namespaces. If you have same-named ingress-resources across different namespaces the error triggers for just one of them, and also resolves after you corrected one of them, leaving you with potentially uncorrected ingress-resources.

openshift_ingress_to_route_controller_ingress_without_class_name 

does not contain a relevant namespace label for the target ingress, just a name. This might be the problem.

That issue is being tracked as OCPBUGS-15253, which is currently in progress with a proposed fix linked.

We are also experiencing this on an OCP 4.12.34 cluster, and are happy to help with any investigation.

For the original problem: On our cluster the alert is only resolved if the ingressClassName is added. Removing the ingress-resource results in a "ghost"-alert that is not closing because "openshift_ingress_to_route_controller_ingress_without_class_name" stays on 1.

We also observed that the alerts are not differentiating between different namespaces. If you have same-named ingress-resources across different namespaces the error triggers for just one of them, and also resolves after you corrected one of them, leaving you with potentially uncorrected ingress-resources.

openshift_ingress_to_route_controller_ingress_without_class_name 

does not contain a relevant namespace label for the target ingress, just a name. This might be the problem.

I did a few initial round of analysis on this, seems like the metric doesn't get updated even when the Ingress is deleted.

» oc get ingress -A
No resources found

As the metrics endpoint still would return the following, despite these ingress(es) being already removed from the cluster:

# HELP openshift_ingress_to_route_controller_ingress_without_class_name Report the number of ingresses that do not specify ingressClassName.
# TYPE openshift_ingress_to_route_controller_ingress_without_class_name gauge
openshift_ingress_to_route_controller_ingress_without_class_name{name="cm-acme-http-solver-2j8vk"} 1
openshift_ingress_to_route_controller_ingress_without_class_name{name="cm-acme-http-solver-bkxsg"} 1
openshift_ingress_to_route_controller_ingress_without_class_name{name="ingress-le-prod"} 1

The expected behaviour would be to atleast set these gauge metric with specific name labels of the ingress to 0 once they're removed.

The metric is created in route-controller-manager, the alert is created in cluster-ingress-operator.  Need to determine where metric can be deleted to make this work.