I'm trying to patch the nginx ingress controller that follows the minikube vm.

Patching is successful using this command:

$ kubectl patch deployment nginx-ingress-controller --type 'json' --namespace kube-system -p '[{"op": "add", "path": "/spec/template/spec/containers/0/args/-", "value": "--profiling"}]'

#-> deployment.extensions/nginx-ingress-controller patched

After patching, the previous state is rolled back automatically. I can see the configuration persisted if I check just after deployment (Like below)

$ kubectl describe deployment/nginx-ingress-controller --namespace kube-system
#--- snip
    Args:
      /nginx-ingress-controller
      --default-backend-service=$(POD_NAMESPACE)/default-http-backend
      --configmap=$(POD_NAMESPACE)/nginx-load-balancer-conf
      --tcp-services-configmap=$(POD_NAMESPACE)/tcp-services
      --udp-services-configmap=$(POD_NAMESPACE)/udp-services
      --annotations-prefix=nginx.ingress.kubernetes.io
      --report-node-internal-ip-address
      --profiling

#---

After rollback the config is reset:

$ kubectl describe deployment/nginx-ingress-controller --namespace kube-system
#--- snip
    Args:
      /nginx-ingress-controller
      --default-backend-service=$(POD_NAMESPACE)/default-http-backend
      --configmap=$(POD_NAMESPACE)/nginx-load-balancer-conf
      --tcp-services-configmap=$(POD_NAMESPACE)/tcp-services
      --udp-services-configmap=$(POD_NAMESPACE)/udp-services
      --annotations-prefix=nginx.ingress.kubernetes.io
      --report-node-internal-ip-address

#---

I cannot see any errors in logs, which should trigger the rollback. The only I can see before the rollback is the deployment triggering shutdown on the pods due to configuration change.