Helm Shows 'Deployed' But App Is Not Working — Fix

# See exactly what resources Helm created
helm get manifest my-release -n production | kubectl apply --dry-run=client -f -
 
# Or list all resources
helm get manifest my-release -n production | grep "^kind:"
 
# Check actual status of those resources
kubectl get all -n production -l app.kubernetes.io/instance=my-release

kubectl describe pod my-app-xxx -n production
# Look for:
# Warning  Failed  ... Failed to pull image "myapp:wrong-tag": not found
# Or: Back-off restarting failed container

# Check what image was deployed
helm get values my-release -n production | grep image
 
# Roll back to last working version
helm rollback my-release -n production
 
# Or upgrade with correct tag
helm upgrade my-release ./chart --set image.tag=v1.2.3 -n production

# Check what values were used
helm get values my-release -n production
 
# Check env vars in the pod
kubectl exec -it my-app-xxx -n production -- env | grep -E "DB|API|SECRET"
 
# Check if secret exists
kubectl get secret my-app-secret -n production
kubectl describe secret my-app-secret -n production

# Override values at deploy time
helm upgrade my-release ./chart \
  --set database.url="postgresql://correct-host:5432/mydb" \
  --set api.key="${API_KEY}" \
  -n production

# Check service endpoints (empty = no pods matching selector)
kubectl get endpoints my-app-service -n production
# NAME             ENDPOINTS
# my-app-service   <none>    ← Problem: no matching pods
 
# Check service selector vs pod labels
kubectl describe service my-app-service -n production | grep Selector
kubectl get pods -n production --show-labels | grep my-app
# Labels need to match selector exactly

# values.yaml
service:
  selector:
    app: my-app    # Must match pod template labels exactly
 
# deployment template
template:
  metadata:
    labels:
      app: my-app   # Must match this
      # app: myapp  ← Typo causes no endpoints

kubectl describe pod my-app-xxx -n production | grep -A5 "Readiness"
# Readiness probe failed: HTTP probe failed with statuscode: 500

# Port-forward and test directly
kubectl port-forward pod/my-app-xxx 8080:8080 -n production
curl http://localhost:8080/health/ready

kubectl patch deployment my-app -n production --type=json \
  -p='[{"op": "remove", "path": "/spec/template/spec/containers/0/readinessProbe"}]'

# Check if CPU is being throttled
kubectl top pods -n production
 
# Check limits
kubectl describe pod my-app-xxx -n production | grep -A6 Limits
 
# Check for throttling events
kubectl describe node $(kubectl get pod my-app-xxx -o jsonpath='{.spec.nodeName}')

helm upgrade my-release ./chart \
  --set resources.limits.cpu=500m \
  --set resources.limits.memory=512Mi \
  -n production

# Check ingress
kubectl describe ingress my-app-ingress -n production
 
# Check ingress controller logs
kubectl logs -n ingress-nginx deployment/ingress-nginx-controller | tail -50
 
# Test directly (bypass ingress)
kubectl port-forward service/my-app-service 8080:80 -n production
curl http://localhost:8080

# Prefix: /api matches /api, /api/v1, /api/users
# Exact: /api only matches /api exactly
spec:
  rules:
    - http:
        paths:
          - path: /
            pathType: Prefix    # Use Prefix for most apps

# 1. Are pods running?
kubectl get pods -n production -l app.kubernetes.io/instance=my-release
 
# 2. Are pods ready?
kubectl get pods -n production | grep "0/"  # Shows not-ready pods
 
# 3. Are there recent events?
kubectl get events -n production --sort-by='.lastTimestamp' | tail -20
 
# 4. Do service endpoints exist?
kubectl get endpoints -n production
 
# 5. Does the app respond directly?
kubectl port-forward service/my-app-service 8080:80 -n production &
curl -v http://localhost:8080/health
 
# 6. What are the actual values deployed?
helm get values my-release -n production