Kubernetes Service Not Reachable: Troubleshooting

Your pods are running, the Service exists, but nothing can reach it. Requests timeout, curl hangs, and your application appears completely disconnected from the network. Service connectivity issues are among the most common Kubernetes problems, and they almost always come down to a few well-known misconfigurations.

Quick Answer

Most Service connectivity issues fall into three categories: selector mismatches, endpoint problems, or DNS failures. Start here:

kubectl get endpoints <service-name>

If the ENDPOINTS column is empty, your Service isn't finding any pods. Check that the Service selector matches the pod labels:

kubectl get svc <service-name> -o jsonpath='{.spec.selector}'
kubectl get pods --show-labels

Step-by-Step Troubleshooting

Step 1: Verify the Service Exists and Has the Right Configuration

Start by inspecting the Service:

kubectl get svc <service-name> -o wide

Check the output for:

For a deeper look:

kubectl describe svc <service-name>

Step 2: Check If the Service Has Endpoints

This is the single most important diagnostic step. If your Service has no endpoints, it cannot route traffic:

kubectl get endpoints <service-name>

You should see pod IPs listed. If the ENDPOINTS column shows <none>, move to Step 3.

For more detail, check the EndpointSlice resources:

kubectl get endpointslices -l kubernetes.io/service-name=<service-name>

Step 3: Diagnose Empty Endpoints

Empty endpoints mean the Service selector doesn't match any running, ready pods. This is the most common cause of unreachable Services.

Compare selectors and labels side by side:

# Get the Service's selector
kubectl get svc <service-name> -o jsonpath='{.spec.selector}' | jq

# List pods with their labels
kubectl get pods --show-labels

Common selector mismatches:

Fix a selector mismatch by either updating the Service or the Deployment labels:

# Option A: Update the Service selector to match existing pods
apiVersion: v1
kind: Service
metadata:
  name: my-service
spec:
  selector:
    app: my-app  # Must match pod template labels exactly
  ports:
  - port: 80
    targetPort: 8080

# Option B: Update the Deployment labels to match the Service
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
spec:
  selector:
    matchLabels:
      app: my-app
  template:
    metadata:
      labels:
        app: my-app  # Must match Service selector
    spec:
      containers:
      - name: my-app
        image: my-app:v1
        ports:
        - containerPort: 8080

Step 4: Verify Pods Are Ready

Even if selectors match, a pod must be in Ready state to receive traffic. Pods that are running but not ready won't appear in the Service endpoints.

kubectl get pods -l <selector-labels>

If pods show 0/1 READY, check why they're not ready:

kubectl describe pod <pod-name>

Look for readiness probe failures in the Events section. A failing readiness probe is a common reason for endpoints disappearing.

Step 5: Check Port Configuration

A frequent but subtle mistake is a mismatch between the Service port, targetPort, and the container's actual listening port:

apiVersion: v1
kind: Service
metadata:
  name: my-service
spec:
  ports:
  - port: 80          # Port the Service listens on
    targetPort: 8080   # Port on the pod to forward to
    protocol: TCP

The targetPort must match the port your container is actually listening on. Verify:

kubectl exec <pod-name> -- ss -tlnp
# or
kubectl exec <pod-name> -- netstat -tlnp

If your container listens on port 3000 but targetPort is set to 8080, traffic will never reach it.

Step 6: Test DNS Resolution

If endpoints exist but the Service is still unreachable by name, the problem might be DNS:

# Run a debug pod to test DNS
kubectl run dns-test --rm -it --image=busybox -- nslookup <service-name>

The Service should resolve to its ClusterIP. Try the fully qualified domain name if the short name doesn't resolve:

nslookup <service-name>.<namespace>.svc.cluster.local

If DNS doesn't resolve at all, check that CoreDNS is running:

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

Step 7: Test Direct Pod Connectivity

To rule out Service-level issues, try connecting to the pod IP directly:

# Get a pod's IP
kubectl get pod <pod-name> -o jsonpath='{.status.podIP}'

# Test connectivity from another pod
kubectl run curl-test --rm -it --image=curlimages/curl -- curl -v http://<pod-ip>:<container-port>

If direct pod access works but the Service doesn't, the problem is in the Service configuration or kube-proxy.

Step 8: Check kube-proxy and Network Policies

If all the above checks pass but the Service is still unreachable:

Verify kube-proxy is running:

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

Check for NetworkPolicies that might block traffic:

kubectl get networkpolicies
kubectl describe networkpolicy <policy-name>

A NetworkPolicy with restrictive ingress rules can silently block Service traffic even when everything else is correctly configured.

Complete Diagnostic Flowchart

Practice These Scenarios

Start the Wrong Selector Challenge →

In this hands-on challenge, you'll:

Start the Expose Internally Challenge →

In this hands-on challenge, you'll:

Start the Partial Outage Challenge →

In this hands-on challenge, you'll:

Prevention Tips