Troubleshoot Network Explorer 🔗
Network Explorer isn’t showing metrics as expected 🔗
When Network Explorer experiences data issues, you might see the following:
Very limited or no TCP metrics and HTTP metrics.
No workloads, or only workloads for
containerd cgroup drivers might be not be the same value.
Check the cgroup drivers of your
containerd to make sure that they match. Both of them have to be either
systemd. For more information, see the Kubernetes documentation on cgroup drivers.
You don’t see Kubernetes metadata in Network Explorer metrics 🔗
Your Network Explorer metrics are not generated with Kubernetes metadata.
Your Kubernetes collector is turned off.
The Kubernetes collector is turned off if you see the following:
k8sCollector: enabled: false
If your Kubernetes collector is turned on, you can determine the root cause based on the logs for the
k8s-relaycontainers in the
Run the following command to find your
kubectl get pods
Run the following commands to retrieve the logs for the containers.
kubectl logs <POD_NAME> -c k8s-watcher kubectl logs <POD_NAME> -c k8s-relay
This is an example error message from the logs. In this case, the service account token is not available.
Error: open /var/run/secrets/kubernetes.io/serviceaccount/token: no such file or directory
On initial startup, the
k8s-watchertries to connect to the
k8s-relay. If the
k8s-relayhas not yet come up, you might see the following set of error messages.
Error: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp [::1]:8712: connect: connection refused"
These messages are expected and you can ignore them.
If your Kubernetes collector is turned off, you need to turn it on by setting
truein the Network Explorer values file.
k8sCollector: enabled: true
For other errors, see the Kubernetes documentation for more information on configuring the service account for the pod to turn on communication with the API server.