(feat/ #374) support proxy timeout

[zjj2wry]

#374
add ingress.kubernetes.io/proxyTimeout annotation in L4-configmap. default alse 600s.

apiVersion: v1
data:
  "99": caicloud-io/serving-manager:8080
kind: ConfigMap
metadata:
  annotations:
    ingress.kubernetes.io/proxyTimeout: 1h
  name: proxy-nginx-tcp
  namespace: kube-system

[k8s-reviewable]

This change is 

[coveralls]

Coverage increased (+0.01%) to 43.965% when pulling 5179323e3bfcd399d6cf88f1486b4ccb7d6cfd87 on zjj2wry:feat/proxy_timeout into faf80ad on kubernetes:master.

[aledbf]

@zjj2wry please use the existing approach to parse annotations https://github.com/kubernetes/ingress/tree/master/core/pkg/ingress/annotations

[aledbf]

Also change the name to something like proxy-stream-timeout

[zjj2wry]

@aledbf now set this config in global. and must restart can be reload.

[coveralls]

Coverage decreased (-0.02%) to 44.993% when pulling 13ab894 on zjj2wry:feat/proxy_timeout into 413b4ad on kubernetes:master.

[aledbf]

/lgtm

[aledbf]

@zjj2wry thanks!

[AEGEGE]

apiVersion: v1
data:
  "2222": kube-ops/jumpserver:2222
kind: ConfigMap
metadata:
  annotations:
    ingress.kubernetes.io/proxyTimeout: 1800s
  labels:
    app: nginx-ingress
    chart: nginx-ingress-1.17.0
    component: controller
    heritage: Tiller
    release: nginx-ingress
  name: nginx-ingress-tcp
  namespace: kube-ops

I set it up like this and restarted all the pods, and found that it hasn't changed. It's still the default 600s.

[AEGEGE]

#374
add ingress.kubernetes.io/proxyTimeout annotation in L4-configmap. default alse 600s.

apiVersion: v1
data:
  "99": caicloud-io/serving-manager:8080
kind: ConfigMap
metadata:
  annotations:
    ingress.kubernetes.io/proxyTimeout: 1h
  name: proxy-nginx-tcp
  namespace: kube-system

apiVersion: v1
data:
"2222": kube-ops/jumpserver:2222
kind: ConfigMap
metadata:
annotations:
ingress.kubernetes.io/proxyTimeout: 1800s
labels:
app: nginx-ingress
chart: nginx-ingress-1.17.0
component: controller
heritage: Tiller
release: nginx-ingress
name: nginx-ingress-tcp
namespace: kube-ops

I set it up like this and restarted all the pods, and found that it hasn't changed. It's still the default 600s.

[whalecold]

#374
add ingress.kubernetes.io/proxyTimeout annotation in L4-configmap. default alse 600s.

apiVersion: v1
data:
  "99": caicloud-io/serving-manager:8080
kind: ConfigMap
metadata:
  annotations:
    ingress.kubernetes.io/proxyTimeout: 1h
  name: proxy-nginx-tcp
  namespace: kube-system

apiVersion: v1
data:
"2222": kube-ops/jumpserver:2222
kind: ConfigMap
metadata:
annotations:
ingress.kubernetes.io/proxyTimeout: 1800s
labels:
app: nginx-ingress
chart: nginx-ingress-1.17.0
component: controller
heritage: Tiller
release: nginx-ingress
name: nginx-ingress-tcp
namespace: kube-ops

I set it up like this and restarted all the pods, and found that it hasn't changed. It's still the default 600s.

try use ingress.kubernetes.io/proxy-stream-timeout.

[taybur43]

Is it solved?
annotations:
ingress.kubernetes.io/proxyTimeout: 1h in tcp service config map but it's still the default 600s