Hi,
first of all thanks for the awesome project!
I’m using Microlink with a Headscale server deployed behind a standard HTTPS reverse proxy (Caddy). As I understand, the current implementation assumes a raw TCP + HTTP/1.1 Upgrade connection to the control-plane server, typically on port 80.
This works in simple direct-access deployments, but it breaks in more common modern setups where:
- The Headscale server is only exposed via HTTPS (port 443)
- Port 80 is closed or not available
- A reverse proxy (e.g. Caddy / Nginx) terminates TLS in front of Headscale
In this setup, the connection fails during the upgrade phase because Microlink is sending a plain HTTP upgrade request to a TLS-enabled endpoint.
Do you plan on adding such a feature in the near term, so that Microlink can operate behind standard HTTPS reverse proxies?
Hi,
first of all thanks for the awesome project!
I’m using Microlink with a Headscale server deployed behind a standard HTTPS reverse proxy (Caddy). As I understand, the current implementation assumes a raw TCP + HTTP/1.1 Upgrade connection to the control-plane server, typically on port 80.
This works in simple direct-access deployments, but it breaks in more common modern setups where:
In this setup, the connection fails during the upgrade phase because Microlink is sending a plain HTTP upgrade request to a TLS-enabled endpoint.
Do you plan on adding such a feature in the near term, so that Microlink can operate behind standard HTTPS reverse proxies?