KongIngress Ignored by Ingress Controller

Kubernetes
1

I am trying to configure a Kong Service to have a longer timeout (the app on the other end takes a log time to process a request, and with the default settings in kong it often times out: “Kong Error: The upstream server is timing out”)

I have the following K8s Ingress:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  annotations:
    configuration.konghq.com: long-timeout
  name: my-app-ingress
  namespace: my-namespace
spec:
  rules:
  - host: example.com
    http:
      paths:
      - backend:
          serviceName: my-app-service
          servicePort: 80
        path: /

Which I try to configure with the following KongIngress object:

apiVersion: configuration.konghq.com/v1
kind: KongIngress
  name: long-timeout
  namespace: my-namespace
proxy:
  connect_timeout: 600000 # 10 minutes

But the kong admin api’s Sevices endpoint shows it has not been updated.
Logs coming from the ingress controller also don’t seem to register that anything has changed. The only log coming out is:
1 kong.go:57] no configuration change, skipping sync to Kong

What am I doing wrong? Is there a special order in which the resources should be created to be registered correctly?
I followed the process describe here, but to no avail.

2

This is a service-level configuration. Hence, you will need to annotate my-app-service resource in k8s with the configuration.konghq.com annotation.