http -> https redirect in Google Kubernetes Engine - kubernetes

Question

I'm looking to redirect all traffic from
http://example.com -> https://example.com like how nearly all websites do.
I've looked at this link with no success:
Kubernetes HTTPS Ingress in Google Container Engine
And have tried the following annotations in my ingress.yaml file.
nginx.ingress.kubernetes.io/configuration-snippet: |
if ($http_x_forwarded_proto != 'https') {
return 301 https://$host$request_uri;
}
nginx.ingress.kubernetes.io/force-ssl-redirect: "true"
nginx.ingress.kubernetes.io/ssl-redirect: "true"
kubernetes.io/ingress.allow-http: "false"
All without any success. To be clear, I can access https://example.com and http://example.com without any errors, I need the http call to redirect to https.
Thanks

Answer 1

GKE uses GCE L7. The rules that you referenced in the example are not supported and the HTTP to HTTPS redirect should be controlled at the application level.
L7 inserts the x-forwarded-proto header that you can use to understand if the frontend traffic came using HTTP or HTTPS. Take a look here: Redirecting HTTP to HTTPS
There is also an example in that link for Nginx (copied for convenience):
# Replace '_' with your hostname.
server_name _;
if ($http_x_forwarded_proto = "http") {
return 301 https://$host$request_uri;
}

Answer 2

Currently, the documentation for how to do this properly (annotations, SSL/HTTPS, health checks etc) is severely lacking, and has been for far too long. I suspect it's because they prefer you to use the App Engine, which is magical but stupidly expensive. For GKE, here's two options:
ingress with a google-managed SSL cert and additional NGINX server configuration in front of your app/site
the NGINX ingress controller with self-managed/third-party SSL certs
The following is steps to a working setup using the former.
1 The door to your app
nginx.conf: (ellipses represent other non-relevant, non-compulsory settings)
user nginx;
worker_processes auto;
events {
worker_connections 1024;
}
http {
...
keepalive_timeout 620s;
## Logging ##
...
## MIME Types ##
...
## Caching ##
...
## Security Headers ##
...
## Compression ##
....
server {
listen 80;
## HTTP Redirect ##
if ($http_x_forwarded_proto = "http") {
return 301 https://[YOUR DOMAIN]$request_uri;
}
location /health/liveness {
access_log off;
default_type text/plain;
return 200 'Server is LIVE!';
}
location /health/readiness {
access_log off;
default_type text/plain;
return 200 'Server is READY!';
}
root /usr/src/app/www;
index index.html index.htm;
server_name [YOUR DOMAIN] www.[YOUR DOMAIN];
location / {
try_files $uri $uri/ /index.html;
}
}
}
NOTE: One serving port only. The global forwarding rule adds the http_x_forwarded_proto header to all traffic that passes through it. Because ALL traffic to your domain is now passing through this rule (remember, one port on the container, service and ingress), this header will (crucially!) always be set. Note the check and redirect above: it only continues with serving if the header value is 'https'. The root and index and location values may differ depending for your project (this is an angular project). keepalive_timeout is set to the value recommended by google. I prefer using the main nginx.conf file, but most people add a custom.conf file to /etc/nginx/conf.d; if you do this just make sure the file is imported into the main nginx.conf http block using an includes statement. The comments highlight where other settings you may want to add when everything is working will go, like gzip/brotli, security headers, where logs are saved, and so on.
Dockerfile:
...
COPY nginx.conf /etc/nginx/nginx.conf
CMD ["nginx", "-g", "daemon off;"]
NOTE: only the final two lines. Specifying an EXPOSE port is unnecessary. COPY replaces the default nginx.conf with the modified one. CMD starts a light server.
2 Create a deployment manifest and apply/create
deployment.yaml:
apiVersion: apps/v1
kind: Deployment
metadata:
name: uber-dp
spec:
replicas: 1
selector:
matchLabels:
app: uber
template:
metadata:
labels:
app: uber
spec:
containers:
- name: uber-ctr
image: gcr.io/uber/beta:v1 // or some other registry
livenessProbe:
failureThreshold: 3
initialDelaySeconds: 60
httpGet:
path: /health/liveness
port: 80
scheme: HTTP
readinessProbe:
failureThreshold: 3
initialDelaySeconds: 30
httpGet:
path: /health/readiness
port: 80
scheme: HTTP
ports:
- containerPort: 80
imagePullPolicy: Always
NOTE: only one specified port is necessary, as we're going to point all (HTTP and HTTPS) traffic to it. For simplicity we're using the same path for liveness and readiness probes; these checks will be dealt with on the NGINX server, but you can and should add checks that probe the health of your app itself (eg a dedicated page that returns a 200 if healthy).The readiness probe will also be picked up by GCE, which by default has its own irremovable health check.
3 Create a service manifest and apply/create
service.yaml:
apiVersion: v1
kind: Service
metadata:
name: uber-svc
labels:
app: uber
spec:
ports:
- name: default-port
port: 80
selector:
app: uber
sessionAffinity: None
type: NodePort
NOTE: default-port specifies port 80 on the container.
4 Get a static IP address
On GCP in the hamburger menu: VPC Network -> External IP Addresses. Convert your auto-generated ephemeral IP or create a new one. Take note of the name and address.
5 Create an SSL cert and default zone
In the hamburger menu: Network Service -> Load Balancing -> click 'advanced menu' -> Certificates -> Create SSL Certificate. Follow the instructions, create or upload a certificate, and take note of the name. Then, from the menu: Cloud DNS -> Create Zone. Following the instructions, create a default zone for your domain. Add a CNAME record with www as the DNS name and your domain as the canonical name. Add an A record with an empty DNS name value and your static IP as the IPV4. Save.
6 Create an ingress manifest and apply/create
ingress.yaml:
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: mypt-ingress
annotations:
kubernetes.io/ingress.global-static-ip-name: [NAME OF YOUR STATIC IP ADDRESS]
kubernetes.io/ingress.allow-http: "true"
ingress.gcp.kubernetes.io/pre-shared-cert: [NAME OF YOUR GOOGLE-MANAGED SSL]
spec:
backend:
serviceName: mypt-svc
servicePort: 80
NOTE: backend property points to the service, which points to the container, which contains your app 'protected' by a server. Annotations connect your app with SSL, and force-allow http for the health checks. Combined, the service and ingress configure the G7 load balancer (combined global forwarding rule, backend and frontend 'services', SSL certs and target proxies etc).
7 Make a cup of tea or something
Everything needs ~10 minutes to configure. Clear cache and test your domain with various browsers (Tor, Opera, Safari, IE etc). Everything will serve over https.
What about the NGINX Ingress Controller?
I've seen discussion it being better because it's cheaper/uses less resources and is more flexible. It isn't cheaper: it requires an additional deployment/workload and service (GCE L4). And you need to do more configuration. Is it more flexible? Yes. But in taking care of most of the work, the first option gives you a more important kind of flexibility — namely allowing you to get on with more pressing matters.

Answer 3

For everyone like me that searches this question about once a month, Google has responded to our requests and is testing HTTP->HTTPS SSL redirection on their load balancers. Their latest answer said it should be in Alpha sometime before the end of January 2020.
Their comment:
Thank you for your patience on this issue. The feature is currently in testing and we expect to enter Alpha phase before the end of January. Our PM team will have an announcement with more details as we get closer to the Alpha launch.
Update: HTTP to HTTPS redirect is now Generally Available: https://cloud.google.com/load-balancing/docs/features#routing_and_traffic_management

Answer 4

GKE uses its own Ingress Controller which does not support forcing https.
That's why you will have to manage NGINX Ingress Controller yourself.
See this post on how to do it on GKE.
Hope it helps.

Answer 5

For what it's worth, I ended up using a reverse proxy in NGINX.
You need to create secrets and sync them into your containers
You need to create a configmap in nginx with your nginx config, as well as a default config that references this additional config file.
Here is my configuration:
worker_processes 1;
events {
worker_connections 1024;
}
http {
default_type application/octet-stream;
# Logging Configs
log_format main '$remote_addr - $remote_user [$time_local] "$request" '
'$status $body_bytes_sent "$http_referer" '
'"$http_user_agent" "$http_x_forwarded_for"';
access_log /var/log/nginx/access.log main;
sendfile on;
keepalive_timeout 65;
# Puntdoctor Proxy Config
include /path/to/config-file.conf;
# PubSub allows 10MB Files. lets allow 11 to give some space
client_max_body_size 11M;
}
Then, the config.conf
server {
listen 80;
server_name example.com;
return 301 https://$host$request_uri;
}
server {
listen 443;
server_name example.com;
ssl_certificate /certs/tls.crt;
ssl_certificate_key /certs/tls.key;
ssl on;
ssl_session_cache builtin:1000 shared:SSL:10m;
ssl_protocols TLSv1 TLSv1.1 TLSv1.2;
ssl_ciphers ECDHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-RSA-RC4-SHA:AES128-GCM-SHA256:HIGH:!RC4:!MD5:!aNULL:!EDH:!CAMELLIA;
ssl_prefer_server_ciphers on;
location / {
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Forwarded-Host $http_host;
# Fix the “It appears that your reverse proxy set up is broken" error.
proxy_pass http://deployment-name:8080/;
proxy_read_timeout 90;
proxy_redirect http://deployment-name:8080/ https://example.com/;
}
}
Create a deployment:
Here are the .yaml files
---
apiVersion: v1
kind: Service
metadata:
name: puntdoctor-lb
spec:
ports:
- name: https
port: 443
targetPort: 443
- name: http
port: 80
targetPort: 80
selector:
app: puntdoctor-nginx-deployment
type: LoadBalancer
loadBalancerIP: 35.195.214.7
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: puntdoctor-nginx-deployment
spec:
replicas: 2
template:
metadata:
labels:
app: puntdoctor-nginx-deployment
spec:
containers:
- name: adcelerate-nginx-proxy
image: nginx:1.13
volumeMounts:
- name: certs
mountPath: /certs/
- name: site-config
mountPath: /etc/site-config/
- name: default-config
mountPath: /etc/nginx/
ports:
- containerPort: 80
name: http
- containerPort: 443
name: https
volumes:
- name: certs
secret:
secretName: nginxsecret
- name: site-config
configMap:
name: nginx-config
- name: default-config
configMap:
name: default
Hope this helps someone solve this issue, thanks for the other 2 answers, they both gave me valuable insight.

Answer 6

As of GKE version 1.18.10-gke.600, you can use FrontendConfig to create HTTP -> HTTPS redirection in Google Kubernetes Engine
HTTP to HTTPS redirects are configured using the redirectToHttps field
in a FrontendConfig custom resource. Redirects are enabled for the
entire Ingress resource so all services referenced by the Ingress will
have HTTPS redirects enabled.
The following FrontendConfig manifest enables HTTP to HTTPS redirects.
Set the spec.redirectToHttps.enabled field to true to enable HTTPS
redirects. The spec.responseCodeName field is optional. If it's
omitted a 301 Moved Permanently redirect is used.
For example:
apiVersion: networking.gke.io/v1beta1
kind: FrontendConfig
metadata:
name: your-frontend-config-name
spec:
redirectToHttps:
enabled: true
responseCodeName: MOVED_PERMANENTLY_DEFAULT
MOVED_PERMANENTLY_DEFAULT is on of the available RESPONSE_CODE field value, to return a 301 redirect response code (default if responseCodeName is unspecified).
You can find more options here: HTTP to HTTPS redirects
Then you have to link your FrontendConfig to the Ingress, like this:
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: your-ingress-name
annotations:
networking.gke.io/v1beta1.FrontendConfig: your-frontend-config-name
spec:
tls:
...