You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have a Contour Ingress controller setup on my Kubernetes instance. I have also setup an Ingress resource together with an HTTProxy resource, per the Contour specification:
apiVersion: projectcontour.io/v1
kind: HTTPProxy
metadata:
name: signalr-proxy
spec:
virtualhost:
fqdn: dev.mycompany.com
tls:
secretName: dev.mycompany.com-tls
routes:
- conditions:
- prefix: /api/notificationhub
enableWebsockets: true # Setting this to true enables websocket for all paths that match /websocket
services:
- name: my-service
port: 80
This configuration is working as expected and when I open the Chrome developer tools for my UI I see the following log on the Console tab:
config vundefined loaded
main.56e2cc4e6f9f283a998a.js:2 [01-003] redirecting...
main.56e2cc4e6f9f283a998a.js:2 Connection to signalR has been established....
2.b4c26156c9f4449d4bbf.js:2 Resetting cache..
In have decided to discard Contour's custom HTTPProxy resource and remain with a plain Ingress resource. To do that I deleted the HTTPProxy resource provided earlier, and on the Ingress resource I added the annotation that enables websocket traffic. The updated Ingress resource is given below:
When I deploy this update I get the following error in the Chrome developer tools console:
config vundefined loaded
main.56e2cc4e6f9f283a998a.js:2 [01-002] redirecting...
main.56e2cc4e6f9f283a998a.js:2 ERROR e
Wn @ main.56e2cc4e6f9f283a998a.js:2
main.56e2cc4e6f9f283a998a.js:2 Connection to signalR has been established....
api/UserSettings/:1 Failed to load resource: the server responded with a status of 405 ()
2.b4c26156c9f4449d4bbf.js:2 Resetting cache...
Visuals:
The Contour deployment has the following set of pods:
One thing I notice is that in the HTTPProxy config, you're matching on a prefix of /api/notificationhub, whereas in the Ingress, you match on /api/notificationhub/ (with trailing slash). Not sure if this matters or not for your service, but I'd try dropping the trailing slash (you'll also need to drop it from the value of the projectcontour.io/websocket-routes annotation) and see if that has any effect.
I have a Contour Ingress controller setup on my Kubernetes instance. I have also setup an Ingress resource together with an HTTProxy resource, per the Contour specification:
Ingress
HTTProxy
This configuration is working as expected and when I open the Chrome developer tools for my UI I see the following log on the Console tab:
In have decided to discard Contour's custom HTTPProxy resource and remain with a plain Ingress resource. To do that I deleted the HTTPProxy resource provided earlier, and on the Ingress resource I added the annotation that enables websocket traffic. The updated Ingress resource is given below:
When I deploy this update I get the following error in the Chrome developer tools console:
Visuals:
The Contour deployment has the following set of pods:
However I have another Ingress resource for Grafana that uses websockets and its working perfectly. The Helm values.yaml file
is provided below:
What am I missing ?
The text was updated successfully, but these errors were encountered: