only ipv6 monitor error

[AEGEGE]

I am using helm to install tigera-operator. Because I need to monitor, I have opened nodeMetricsPort: 9091, but in the only ipv6 environment, address="fddd:3bcc:a689::66" is found in the Input Labels captured by prometheus, resulting in The monitoring rule failed. Normally it should be address="[fddd:3bcc:a689::66]:$port", so I hope to inject a port configuration into it:

    ports:
         - containerPort: 9091
           hostPort: 9091
           protocol:TCP

But the injection failed. I know it may be a problem with my configuration, but I can't find an example. My wrong configuration is:

installation:
   registry: aetest.com:5000
   calicoNetwork:
     ipPools:
     -blockSize: 122
       cidr: fcaa:8888::/64
       encapsulation: VXLAN
       natOutgoing: Enabled
       nodeSelector: all()
     mtu: 0
     nodeAddressAutodetectionV6:
       interface: eth.*|en.*
   enabled: true
   kubeletVolumePluginPath: /data/k8sbasedata/kubelet
   kubernetesProvider: ""
   nodeMetricsPort: 9091
   typhaMetricsPort: 9093
   CalicoNodeDaemonSet:
     template:
       containers:
         ports:
         - containerPort: 9091
           hostPort: 9091
           protocol:TCP

[caseydavenport]

Hm, setting that field should enable the following annotations on the calico/node DaemonSet:

https://github.com/tigera/operator/blob/e5880ff2edf627c98cf12ad29f24bcab011be78d/pkg/render/node.go#L906-L909

I would expect Prometheus to pick up on those, but perhaps we also need to add the containerPort to the spec when that happens?

Right now, the only options available for CalicoNodeDaemonSet would be the resource requests and requirements, per the API: https://github.com/tigera/operator/blob/e5880ff2edf627c98cf12ad29f24bcab011be78d/api/v1/calico_node_types.go#L25-L37

Likely not the problem, but CalicoNodeDaemonSet: should be lower case: calicoNodeDaemonSet

[AEGEGE]

Hm, setting that field should enable the following annotations on the calico/node DaemonSet:

https://github.com/tigera/operator/blob/e5880ff2edf627c98cf12ad29f24bcab011be78d/pkg/render/node.go#L906-L909

I would expect Prometheus to pick up on those, but perhaps we also need to add the containerPort to the spec when that happens?

Right now, the only options available for CalicoNodeDaemonSet would be the resource requests and requirements, per the API: https://github.com/tigera/operator/blob/e5880ff2edf627c98cf12ad29f24bcab011be78d/api/v1/calico_node_types.go#L25-L37

Likely not the problem, but CalicoNodeDaemonSet: should be lower case: calicoNodeDaemonSet

In the current ipv6 environment, if you add the containerPort, the Input Labels are address="[fddd:3bcc:a689::66]:$port". prometheus's relabel_configs can capture and replace it

      - action: replace
        regex: ([^:]+)(?::\d+)?;(\d+)
        replacement: $1:$2
        source_labels:
        - __address__
        - __meta_kubernetes_pod_annotation_prometheus_io_port
        target_label: __address__
      - action: replace
        regex: (\[[^]]+\])(?::\d+)?;(\d+)
        replacement: $1:$2
        source_labels:
        - __address__
        - __meta_kubernetes_pod_annotation_prometheus_io_port
        target_label: __address__

At present, I use a less rigorous relabel_configs and it can meet the usage requirements, but I don’t know if there will be problems in the future.

      - action: replace
        regex: '^([a-fA-F0-9:]{2,})$'
        replacement: '[$1]'
        source_labels:
        - __address__
      - action: replace
        regex: ([^:]+)(?::\d+)?;(\d+)
        replacement: $1:$2
        source_labels:
        - __address__
        - __meta_kubernetes_pod_annotation_prometheus_io_port
        target_label: __address__
      - action: replace
        regex: (\[[^]]+\])(?::\d+)?;(\d+)
        replacement: $1:$2
        source_labels:
        - __address__
        - __meta_kubernetes_pod_annotation_prometheus_io_port
        target_label: __address__

So I think if you can add the containerPort, it is the simplest solution.