WebSocket keepalive/pings

@jan-law Try adding these annotations to each Ingress:

nginx.ingress.kubernetes.io/proxy-read-timeout: 3600
nginx.ingress.kubernetes.io/proxy-send-timeout: 3600

This seems to be the recommended solution from the NGINX Ingress documentation: https://kubernetes.github.io/ingress-nginx/user-guide/miscellaneous/#websockets

_Originally posted by @ebaron in https://github.com/cryostatio/cryostat-operator/issues/279#issuecomment-983039470_

[ebaron]

Since these annotations apply only to Kubernetes, what would be the OpenShift equivalent of setting the nginx timeouts?

I found this in the OpenShift documentation: https://docs.openshift.com/container-platform/4.9/networking/routes/route-configuration.html Variable Default Description ROUTER_DEFAULT_TUNNEL_TIMEOUT 1h Length of time for TCP or WebSocket connections to remain open. This timeout period resets whenever HAProxy reloads.

So it would be an hour by default, but I don’t see an annotation to modify it when creating the Route.

For future reference: there does appear to be annotation for OpenShift routes: haproxy.router.openshift.io/timeout-tunnel [1]

[1] https://docs.openshift.com/container-platform/4.9/networking/routes/route-configuration.html#nw-route-specific-annotations_route-configuration

[andrewazores]

Since these annotations apply only to Kubernetes, what would be the OpenShift equivalent of setting the nginx timeouts?

I found this in the OpenShift documentation: https://docs.openshift.com/container-platform/4.9/networking/routes/route-configuration.html Variable Default Description ROUTER_DEFAULT_TUNNEL_TIMEOUT 1h Length of time for TCP or WebSocket connections to remain open. This timeout period resets whenever HAProxy reloads.

So it would be an hour by default, but I don’t see an annotation to modify it when creating the Route.

That would definitely explain why this behaviour shows up so easily in minikube and not in CRC!