Server stops functioning when rustdesk.com license server is unreachable

Bug Description

Our self-hosted RustDesk Server Pro (v1.7.5) stopped functioning when rustdesk.com became unreachable. Both the client app connections and the web console went down simultaneously. No configuration changes were made on our side — the server recovered on its own once rustdesk.com came back online.

The hbbs log shows the license check failing:

ERROR [src/lic.rs:207] network error: reqwest::Error { kind: Request, url: "https://rustdesk.com/api/lic/license/check", source: hyper_util::client::legacy::Error(Connect, ConnectError("dns error", Custom { kind: Uncategorized, error: "failed to lookup address information: Temporary failure in name resolution" })) }

Additionally, the rustdesk client app reported it could not reach admin.rustdesk.com.

Impact

• All client heartbeats stopped working
• Web console became inaccessible
• Remote desktop connections could not be established
• A fully self-hosted deployment was rendered unusable due to an external dependency

Expected Behavior

A self-hosted server with a valid, already-activated license should continue to operate when rustdesk.com is temporarily unreachable. Suggestions:

• Cache the license locally with a grace period (e.g. 7-30 days) so the server continues operating during outages
• Degrade gracefully — log a warning but don't stop serving clients
• Document the dependency — the current docs do not mention that rustdesk.com availability is required for ongoing operation

Environment

• RustDesk Server Pro 1.7.5
• Debian (self-hosted, bare metal)
• License was valid and previously verified (License OK on prior startups)

Related

• #894 — License gating prevents admin actions even during setup

Start in src/lic.rs around the reported line 207 and trace the license-check flow and its effect on serving clients and the web console. Read related issue #894 for context. Done means a valid, previously activated self-hosted license continues operating during a temporary rustdesk.com outage, with the expected behavior and dependency documented.