knot knot

Tunnels

Knot tunnels allow HTTP and HTTPS services running on a local workstation or within a space to be exposed on the internet via a Knot server.

Configuring the Server

Tunnels require the server’s tunnel listener to be configured — see Tunnel Server Configuration for the listen_tunnel and tunnel_domain settings and their DNS requirements.

Tunnel addresses are built as <user>--<tunnel name>.<tunnel domain>, for example example--tunnel1.tunnel.knot.internal.

Creating a Tunnel

Tunnels can be created either on a local machine (desktop) or within a space (managed by the knot agent).

  • Desktop tunnels (described below) run as a foreground process on your local machine and live for the life of that command.
  • Agent tunnels, started inside a space, can optionally be handed to the knot agent so they keep running after the launching command exits. See Agent Tunnels for --daemon, stop, and list.

Creating a Tunnel on a Local Machine

  1. Connect to the Knot Server Open a terminal and connect to the Knot server by running the following command (replace the URL with the actual server address):

    knot connect https://knot.internal:3000
    • The connect command only needs to be run once or if the login expires.

  2. Create a Tunnel Run the tunnel command with the protocol, port, and tunnel name:

    knot tunnel http 8080 test1
    • This creates a tunnel to the local port 8080 with the address example--test1.tunnel.knot.internal.
    • Stopping the knot tunnel command will stop and remove the tunnel.

Creating a Tunnel in a Space

Run the same command from a terminal inside a space:

knot tunnel http 8080 test1

By default this behaves like the desktop version — the tunnel lives for the life of the command. To let the knot agent own the tunnel so it keeps running after the command exits, add --daemon, and use knot tunnel stop / knot tunnel list to manage it. See Agent Tunnels for the full agent-managed workflow.