How are the multi agent secure the connections between the agents?

[ngocphamm]

Hi,

I don't see a technical detail post about this (yet), so I'm asking here.

For Portainer, they have this as the explanation of how the feature works. https://portal.portainer.io/knowledge/how-does-portainer-secure-connectivity-to-and-from-agents-and-edge-agents

In the Pull request for the feature, I saw the mention that the approach is similar, but how do Dockge make sure it's the only one connecting to the remote agent? The remote agent will have to expose itself via the port 5001 by default, I think, and it's available for anyone who looks, right?