Solution: Ingress Misroute

Root Cause

The Ingress resource references a service named escape-svc, but the actual service is named escape-service:

# In the Ingress (WRONG):
backend:
  service:
    name: escape-svc      # Typo! Missing "-ervice"
    port:
      number: 80

# Actual service name:
metadata:
  name: escape-service    # Correct name

When the Ingress controller tries to route traffic to escape-svc, it can't find the service, resulting in a 404 or 503 error depending on the ingress controller.

Diagnosis Steps

# Step 1: Verify pod and service are working
kubectl get pods,svc -n escape-room-bad-gateway
# Pod is Running, Service exists

# Step 2: Check the Ingress
kubectl get ingress -n escape-room-bad-gateway
kubectl describe ingress escape-ingress -n escape-room-bad-gateway
# Look at the "Rules" section and "Backends"

# Step 3: Compare service names
kubectl get svc -n escape-room-bad-gateway -o name
# Output: service/escape-service

kubectl get ingress escape-ingress -n escape-room-bad-gateway -o jsonpath='{.spec.rules[0].http.paths[0].backend.service.name}'
# Output: escape-svc  ← MISMATCH!

# Step 4: Test internal connectivity (should work)
kubectl run test-curl --rm -it --image=curlimages/curl --restart=Never \
  -n escape-room-bad-gateway -- curl -s http://escape-service
# Returns nginx welcome page - service works!

The Fix

Option 1: Patch the Ingress

kubectl patch ingress escape-ingress -n escape-room-bad-gateway \
  --type='json' \
  -p='[{"op": "replace", "path": "/spec/rules/0/http/paths/0/backend/service/name", "value": "escape-service"}]'

Option 2: Edit the Ingress Directly

kubectl edit ingress escape-ingress -n escape-room-bad-gateway

Change:

backend:
  service:
    name: escape-svc

To:

backend:
  service:
    name: escape-service

Option 3: Replace via YAML

kubectl get ingress escape-ingress -n escape-room-bad-gateway -o yaml | \
  sed 's/escape-svc/escape-service/g' | \
  kubectl apply -f -

Verification

# Check the Ingress now references the correct service
kubectl describe ingress escape-ingress -n escape-room-bad-gateway

# If you have the ingress controller and /etc/hosts configured:
curl -H "Host: escape.local" http://localhost/api

# Or use port-forward to the ingress controller and test
# Or run a curl pod with the Host header:
kubectl run test-ingress --rm -it --image=curlimages/curl --restart=Never \
  -n escape-room-bad-gateway -- \
  curl -s -H "Host: escape.local" http://ingress-nginx-controller.ingress-nginx/api

Lessons Learned

1. Service names must match exactly - typos cause routing failures
2. Ingress issues often result in 404/503 errors, not pod failures
3. Use kubectl describe ingress to see the resolved backends
4. Always verify the service name exists before referencing it in an Ingress
5. Internal connectivity (Service) can work while external (Ingress) fails

Real-World Considerations

This commonly happens when:

• Copy-pasting Ingress configs and forgetting to update service names
• Service renamed but Ingress not updated
• Different naming conventions across teams (svc vs service)
• Auto-generated names don't match manual Ingress configs

Best practices:

• Use Helm/Kustomize to template service names consistently
• Validate Ingress backends exist before deployment
• Monitor Ingress controller logs for routing errors
• Use meaningful, consistent naming conventions
• Consider using ExternalName services for external dependencies
• Test Ingress rules in staging before production