Kubernetes Service Not Found (DNS) Fix

In this tutorial, you'll learn about Kubernetes Service Not Found (DNS) Fix. We cover key concepts, practical examples, and best practices.

The Problem

A pod cannot reach another service by name:

curl: (6) Could not resolve host: my-service

Or in logs:

Error: getaddrinfo ENOTFOUND my-service

The nslookup inside the pod fails:

kubectl exec my-pod -- nslookup my-service
# ** server can't find my-service: NXDOMAIN

Kubernetes DNS resolution relies on CoreDNS (or kube-dns) running in the cluster. When DNS fails, services are unreachable by name.

Quick Fix

Step 1: Check CoreDNS pods are running

kubectl get pods -n kube-system -l k8s-app=kube-dns

NAME                       READY   STATUS    RESTARTS
coredns-7d6c9b5d8f-abc12   1/1     Running   0
coredns-7d6c9b5d8f-xyz99   1/1     Running   0

If CoreDNS pods are not running, describe them:

kubectl describe pods -n kube-system -l k8s-app=kube-dns

Step 2: Check CoreDNS logs

kubectl logs -n kube-system -l k8s-app=kube-dns

Look for errors like:

[ERROR] plugin/errors: 2 example.com. A: dial udp 8.8.8.8:53: i/o timeout

This means CoreDNS cannot reach upstream DNS servers. Check the CoreDNS ConfigMap:

kubectl get configmap coredns -n kube-system -o yaml

WRONG -- incorrect upstream DNS:

Corefile: |
  .:53 {
      kubernetes cluster.local in-addr.arpa ip6.arpa {
          pods insecure
          fallthrough in-addr.arpa ip6.arpa
      }
      forward . 8.8.8.8  # wrong if cluster has no internet access
  }

RIGHT -- use cluster DNS or a known working resolver:

Corefile: |
  .:53 {
      kubernetes cluster.local in-addr.arpa ip6.arpa {
          pods insecure
          fallthrough in-addr.arpa ip6.arpa
      }
      forward . /etc/resolv.conf
  }

Step 3: Verify the service exists in the correct namespace

kubectl get svc --all-namespaces | grep my-service

Services are only resolvable within the same namespace by short name:

WRONG -- using short name across namespaces:

curl http://my-service  # fails if my-service is in a different namespace

RIGHT -- use the fully qualified name:

curl http://my-service.other-ns.svc.cluster.local

Or use the namespace-qualified short name:

curl http://my-service.other-ns

Step 4: Check pod DNS policy

kubectl get pod my-pod -o jsonpath='{.spec.dnsPolicy}'

WRONG -- None without a custom config:

dnsPolicy: None  # only works if you provide dnsConfig

RIGHT -- use ClusterFirst (default):

dnsPolicy: ClusterFirst

For host-networked pods, use:

dnsPolicy: ClusterFirstWithHostNet

Step 5: Test DNS resolution from a debug pod

kubectl run dns-test --image=busybox -it --rm --restart=Never -- nslookup kubernetes.default.svc.cluster.local

Expected output:

Server:    10.96.0.10
Address 1: 10.96.0.10 kube-dns.kube-system.svc.cluster.local

Name:      kubernetes.default.svc.cluster.local
Address 1: 10.96.0.1

If this fails, CoreDNS is not serving DNS for the cluster domain. Increase CoreDNS replicas or check for network policies blocking DNS traffic on port 53.

Use DodaTech's DNS Inspector to run DNS queries against your cluster and visualize the resolution chain from pod to CoreDNS to upstream.

Prevention

• Always use FQDN (service.namespace.svc.cluster.local) for cross-namespace communication.
• Monitor CoreDNS metrics (requests, errors, latency) with Prometheus.
• Run at least two CoreDNS replicas for high availability.
• Configure CoreDNS resource requests to prevent OOM kills.
• Test DNS resolution as part of your deployment pipeline.

Common Mistakes with service not found

1. Overlapping type class instances that cause GHC to reject the program with ambiguous dispatch errors
2. Non-exhaustive pattern matches that compile with warnings then crash at runtime
3. Misunderstanding that String is [Char] with poor performance for large text operations

These mistakes appear frequently in real-world K8S code. DodaTech's contributors have identified these patterns through analysis of open-source projects and production systems.

Practice Exercise

Write a pure function that safely divides two integers using Maybe, then test it with edge cases like division by zero and negative numbers.

This exercise reinforces the concepts covered in this guide. Try implementing it before checking online solutions.

FAQ

hosts {
  192.168.1.50 my-custom-service.local
  fallthrough
}