Restarting tsdproxy container makes new devices on Tailscale network with a suffix -1

[IoannisPant]

I am using tsdproxy on a Synology NAS to provide Tailscale DNS names for multiple Docker containers.
While the setup works well, I have encountered repeated issues with duplicate machine entries (e.g., immich + immich-1, portainer + portainer-1) being created in my Tailscale tailnet.

This happens when the tsdproxy container restarts or after a Tailscale service outage.

When tsdproxy reconnects to Tailscale, it sometimes registers a new machine ID instead of reusing the existing one.

This results in duplicate device entries in the Machines list (*-1, *-2, etc.).

These duplicate machines remain until manually deleted.

The issue persists when I am using a long-lived auth key or rotating auth keys via API.

I would like to know:

Whether tsdproxy is expected to persist and reuse its machine key between restarts.

If there is a recommended way to:

persist the Tailscale state directory,

ensure stable machine identities,

or prevent *-1 / duplicate nodes from being created.

Any guidance or clarification on how tsdproxy should handle machine key persistence would be appreciated.

Thank you !