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

For information on the advisory, 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-2022:7399

This bug fix does not require a release note as the bug was in a new feature that has not shipped.

rh-ee-arsen Yes, I also noticed route Ingress status wasn't changed after the namespace label was changed. Let's wait the known issue fixed and this metric issue will gone automatically. Thanks.

rh-ee-arsen yes, in step 8, the two names are for two different things. It would be great changing the "name" to "shard", thanks.

shudili@redhat.com the issue with change of namespace label is a known one. This is due to the fact that the route Ingress status also does not get updated when the namespace label is changed. This is tracked in https://issues.redhat.com/browse/OCPBUGS-1689. The fix for this bug will also fix the issue of updating the metric value as the metric value is dependent on the route Ingress status.

shudili@redhat.com in step 8 the first column gives the name of the metric, the rest of the columns apart from the last one gives the labels of the metric and the last column gives the corresponding value for the metric. As one label showing the name of the shard is called 'name', that's why 2 columns are having the same name. The label 'name' can be changed to 'shard' to avoid the confusion.

I did the test of Namespace labelselector in an ingress-controller: After I add the label to a namespace(test), the route_metrics_controller_routes_per_shard stats for this ingress-controller will be increased as expected. But after I remove the label from the namespace(test), the stats isn't changed, which should be decreased by the number of routes in the namespace(test).

1. There are totally 18 routes in all namespaces, namespace test has 6 routes(4 routes has "shard: alpha" label), namespace test2 has 4 routes(namespace test2 has "type: sharded" label)
2. There are totally 4 ingress-controllers, ic internal2 has routeSelector("shard: alpha" label), ic extertest3 has namespaceSelector("type: sharded" label)
3. check the stats of "route_metrics_controller_routes_per_shard", as shows in below as expected:
   name                           value
   default                      18
   extertest3                  4
   internal1                    18      
   internal2                    4
4.  add "type: sharded" label to namespace test which has 6 routes, check the stats, it is also as expected
   name                           value
   default                      18
   extertest3                  10
   internal1                    18      
   internal2                    4
5.  remove "type: sharded" label from namespace test, expect the stats for extertest3 changes back to 4, but it is still 10
6. "type: sharded" label is removed from ns test

% oc get namespace test -oyaml | grep labels: -A5

  labels:

    kubernetes.io/metadata.name: test

    pod-security.kubernetes.io/enforce: privileged

    pod-security.kubernetes.io/enforce-version: v1.24

    security.openshift.io/scc.podSecurityLabelSync: "false"

  name: test

% 

1.  check one route's status under ns test 

% oc get route -n test --no-headers=true | wc -l

       6

%

% oc get route -n test edge11 -oyaml
...

status:

  ingress:

  - conditions:

    - lastTransitionTime: "2022-10-26T06:09:46Z"

      status: "True"

      type: Admitted

    host: edge11-test1.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-default.apps.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: default

    wildcardPolicy: None

  - conditions:

    - lastTransitionTime: "2022-10-26T06:09:46Z"

      status: "True"

      type: Admitted

    host: edge11-test1.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-internal1.internal1.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: internal1

    wildcardPolicy: None

  - conditions:

    - lastTransitionTime: "2022-10-26T06:14:15Z"

      status: "True"

      type: Admitted

    host: edge11-test1.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-internal2.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: internal2

    wildcardPolicy: None

  - conditions:

    - lastTransitionTime: "2022-10-26T07:06:10Z"

      status: "True"

      type: Admitted

    host: edge11-test1.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-extertest3.extertest3.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: extertest3

    wildcardPolicy: None

%

1. check the stats again
   name                           value
   default                      18
   extertest3                  10
   internal1                    18      
   internal2                    4

rh-ee-arsen From the route's status, the stats of "route_metrics_controller_routes_per_shard" were correct, thanks again for the explanation. So please take a look at step 8 if it would be necessary to fix the minor name issue, thanks!
8. Also there was a minor issue: As the attached picture showed, there were two name in the header line

Name name value
route_metrics_controller_routes_per_shard   default 15
route_metrics_controller_routes_per_shard   extertest3 15
route_metrics_controller_routes_per_shard   internal1 15
route_metrics_controller_routes_per_shard   internal2 1

rh-ee-arsen Thanks for the explanation. I had thought cluster's domain would work for the routes.

I just create two custom ingress-controllers and two routes, and put one route's status here

status:

  ingress:

  - conditions:

    - lastTransitionTime: "2022-10-26T02:24:04Z"

      status: "True"

      type: Admitted

    host: int1edge1-test.internal1.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-internal2.internal2.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: internal2

    wildcardPolicy: None

  - conditions:

    - lastTransitionTime: "2022-10-26T02:24:04Z"

      status: "True"

      type: Admitted

    host: int1edge1-test.internal1.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-default.apps.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: default

    wildcardPolicy: None

  - conditions:

    - lastTransitionTime: "2022-10-26T02:24:04Z"

      status: "True"

      type: Admitted

    host: int1edge1-test.internal1.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerCanonicalHostname: router-internal1.internal1.shudi-412gcpkf2.qe.gcp.devcluster.openshift.com

    routerName: internal1

    wildcardPolicy: None

mmasters1@redhat.com Yes, I didn't configure any selectors for extertest3, and internal1.