Select the protocol to be used by Signalr

[n2029-ndensan]

Is there an existing issue for this?

• I have searched the existing issues

Is your feature request related to a problem? Please describe the problem.

Connecting clients fail when scaling in during autoscaling.

Describe the solution you'd like

We want to be able to select the protocol to be used by Signalr, and by selecting long polling, we want to be able to scale in when autoscaling. Connecting users without errors.

Additional context

We plan to use Blazor Server. One of the challenges is that it is difficult to scale in.

[BrennanConroy]

Connecting clients fail when scaling in during autoscaling.

So when reducing machines you get connection failures, presumably because the client(s) were connected to the machines that are no longer provisioned? That's not really something SignalR can do anything about.

by selecting long polling, we want to be able to scale in when autoscaling

I don't see the relationship here. Why is long polling important to autoscaling?

[n2029-ndensan]

If it is not a problem with the protocol, is there a problem with the implementation of SignalR on the application side? The state is maintained in memory on the server side. If it is retained, it cannot be stored in external resources such as DB or Redis.If the state is retained in server memory and it takes time to change the SignalR implementation, is it possible to monitor whether SignalR is connected? With AzureMonitor etc.If connection monitoring is possible, I think it would be possible to temporarily disconnect SignalR with SSR and scale in when there is no SignalR connection.