Stef Walter

This is fascinating and really cool. But it does cause me to wonder about the use cases? Does this mean that the cockpit-bridge must stay running throughout the entirety of...

So it appears there is one very fundamental limitation. The cockpit protocol operates across a tree. A tree that has directionality: * Inbound (ie: toward the leaves, "bridges", the various...

Perhaps implementing this at the CockpitRouter level would help avoid this limitation? It appears that would mean inspecting outbound messages though.

You're right. That's what I get for doing code review before drinking my Coffee. :frowning_face:

Hi @lenticularis39 ... sorry for the delay. There was one aspect of this that was missing. We should treat redirect channels that do not (or no longer) exist as closed...

I'm trying to understand the security implications of this. There is a bit of a caveat where connections to different hosts are not strictly part of the same login session,...

That's correct, although it encourages behavior that we may later prohibit. Given: Web Browser -> Host A -> Host B Currently we don't encourage Host B to perform actions to...

> The first question is whether multi-machine operations (as mentioned in #14560; like the aforementioned Cockpit container migration, a file manager supporting copying files between machines and so on) are...

Interesting list of features. Seems that they would help round out that functionality nicely. Do you have plans to work on any parts of this? If so, I can help...

Good point. On the off chance: Do you have a pod/container that exhibits this behavior? That would make it easier for someone to get started and develop a solution.