k8s Flashcards
← Back to blog
CoreDNSNetworkingCKATroubleshooting

CoreDNS Down: When Kubernetes DNS Breaks

nslookup fails inside pods but apps look fine? Cluster DNS is a platform service — check kube-system before you restart application Deployments.

1 min read

A microservice logs:

dial tcp: lookup redis.cache.svc.cluster.local: i/o timeout

Redis pod is Running. The Service exists. Still — DNS is the failure mode.

Quick in-cluster test

kubectl run dns-test --rm -it --image=busybox:1.36 -- nslookup kubernetes.default

If that fails, the problem is cluster DNS, not your app manifest.

Check CoreDNS

kubectl get pods -n kube-system -l k8s-app=kube-dns
kubectl logs -n kube-system -l k8s-app=kube-dns --tail=50

CrashLoop, resource starvation, or ConfigMap loop misconfiguration (/etc/coredns/Corefile) are frequent causes.

Also verify:

kubectl get svc -n kube-system kube-dns

ClusterIP should match /etc/resolv.conf in application pods (nameserver 10.96.0.10 varies by cluster).

CKA vs CKAD lens

RoleTypical fix depth
**CKAD**Recognize DNS failure; fix app ConfigMap upstream URL if wrong; know to check CoreDNS exists
**CKA**Repair CoreDNS Deployment, Corefile, kube-proxy/CNI interaction, node DNSPolicy

On CKA you may SSH to nodes; on CKAD you usually fix workload config after confirming DNS is broken cluster-wide.

Do not

  • Roll every Deployment in the namespace (DNS still broken)
  • Point apps at hard-coded ClusterIPs (fragile; does not fix CoreDNS)
  • Assume external DNS — in-cluster names need CoreDNS

CKA Exam Prep includes cluster networking scenarios; Service Connectivity covers app-side DNS/tracing.