KubernetesMid

Can't resolve a Service name inside the cluster

If a pod can't resolve a Service name, test with nslookup from inside the cluster, then check resolv.conf, CoreDNS pods, logs and endpoints.

What you see

  • wget: unable to resolve host address 'hostnames'
  • Resolving hostnames (hostnames)... failed: Name or service not known.
  • nslookup: can't resolve 'kubernetes.default'

What it means

Pods find Services by DNS name. The cluster DNS (CoreDNS) answers those names. When a name fails, either the Service does not exist, or the pod cannot reach working cluster DNS.

How to confirm the cause

  1. 1
    Does the Service exist?
    kubectl get svc hostnames

    Example from the official docs

    No resources found.
    Error from server (NotFound): services "hostnames" not found

    This is what a missing Service looks like in the docs.

  2. 2
    Try the name from a throwaway pod
    kubectl run -it --rm --restart=Never busybox --image=registry.k8s.io/busybox:1.27.2 sh

    Then run nslookup hostnames. If that fails, try hostnames.default, then hostnames.default.svc.cluster.local.

  3. 3
    Test the cluster DNS itself
    kubectl exec -i -t dnsutils -- nslookup kubernetes.default

    Example from the official docs

    Server:    10.0.0.10
    Address 1: 10.0.0.10
    
    Name:      kubernetes.default
    Address 1: 10.0.0.1

    Create the dnsutils pod first with kubectl apply -f https://k8s.io/examples/admin/dns/dnsutils.yaml.

  4. 4
    Check the pod's DNS settings
    kubectl exec -ti dnsutils -- cat /etc/resolv.conf

    Example from the official docs

    search default.svc.cluster.local svc.cluster.local cluster.local
    nameserver 10.0.0.10
    options ndots:5

    The search line can differ by cloud provider.

  5. 5
    Are the CoreDNS pods up?
    kubectl get pods --namespace=kube-system -l k8s-app=kube-dns
  6. 6
    Read the CoreDNS logs
    kubectl logs --namespace=kube-system -l k8s-app=kube-dns

    Look for anything suspicious or unexpected.

  7. 7
    Does the DNS Service have endpoints?
    kubectl get endpointslice -l kubernetes.io/service-name=kube-dns --namespace=kube-system

Causes and fixes

  • The Service does not exist or is in another namespace

    HOW TO TELL
    kubectl get svc returns NotFound.
    FIX
    Create the Service, for example kubectl expose deployment hostnames --port=80 --target-port=9376, or use the namespace-qualified name.
  • CoreDNS is not running

    HOW TO TELL
    No pod, or a failed or completed pod, for k8s-app=kube-dns.
    FIX
    The docs say the DNS add-on may not be deployed by default in your environment and may need to be deployed manually.
  • DNS Service has no endpoints

    HOW TO TELL
    The kube-dns EndpointSlice is empty.
    FIX
    Follow the Service endpoints runbook for the kube-dns Service in kube-system.

How to stop it coming back

  • Include a DNS check in your cluster readiness tests.
  • Alert when CoreDNS pods are not Ready.

Keep going

Sources

Kubernetes documentation (kubernetes.io), read 18 Nov 2026. The docs version selector listed v1.33 to v1.37.

Search

Search cheat sheets and projects