Decide whether Tailscale Serve should survive a restart

Background sharing can return automatically after reboot; foreground sharing requires another start.

Source facts

Tailscale Serve shares a local service within a tailnet and can act as a reverse proxy, file server, or TCP forwarder. Its command reference distinguishes foreground operation from background operation. With the background flag, Serve persists until disabled and automatically resumes after the device reboots or Tailscale restarts. Without that flag, an operator must start Serve again after a restart. When used with Tailscale Services, background operation is the default. Serve provides status and configuration-management commands for inspecting the sharing arrangement.

Source: Tailscale — Serve command.

Applicability

This applies to the current Serve CLI documented after the command changes introduced in Tailscale 1.52. Check the installed client and platform limitations before reusing an older sharing procedure.

DSE recommendation

DSE recommends choosing persistence explicitly for each shared service. A temporary troubleshooting page should have an owner and removal deadline; a supported internal application should have startup and recovery expectations. Record the backend, listener, path, and permitted audience in the change record. Before restarting a device, inspect its Serve configuration and decide whether automatic resumption is desired. Include the sharing configuration in application retirement so an old diagnostic endpoint does not quietly return with the next reboot.

Verification

On a nonproduction device, test the selected restart behavior and inspect Serve status afterward. Confirm that the intended resource, rather than an unintended directory or backend path, is available. Disable a temporary share and repeat the restart check to verify that it remains removed.

Official reference

Tailscale — Serve command. Reviewed October 15, 2026.