V8 Redundancy for client end

First time doing scenario with public facing redundancy on v8 with perspective, and I have yet to figure out how the client end can fail over or switch to the backup’s address.

For years using the Vision module the backup server location was to be found in the .jnlp file, and was configured within Ignition. I have encountered with doing redundant servers with mqtt broker that the primary shuts down the listening port when the role is transferred to the the backup. When using HAProxy this would signal one of the servers is down, and the other would then be up so connections would be handled through the proxy to the backup.

But Perspective would not work this way. So I would like to know what others, if any, have done to have clients connect to the backup during a fail over without having to manually connect to a different address

I would expect first-class support for redundancy to only happen in Perspective Workstation, not in browsers. (Haven’t tried…)

The perspective session is a ‘single page app’ - the only critical network request is the initial payload. If the persistent websocket to the backend goes down for some reason, the frontend session is locally intelligent enough to “fail over” to the backup gateway (as long as it already knows about it).

At least, in theory.