Server Version#: Synology DSM 7.4, Plex Media Server 1.43.3.10861
Player Version#: iOS 2026.16.0 (issue also present in 2026.14.0)
Device: iPhone 14pro, iOS 26.6
Setup:
- Plex Media Server runs on a Synology NAS with a direct public IP (no router NAT)
- External access is provided via a reverse proxy (Synology DSM reverse proxy, HTTPS, on a custom non-standard port) forwarding to localhost:32400
- Port 32400 is blocked at the firewall for all external IPs (only reachable through the reverse proxy domain + custom port)
- Plex Remote Access is intentionally disabled in server settings
- A custom server access URL (https://mydomain.example.com:XXXXX) is registered under Settings > Network > “Custom server access URLs” so plex.tv can advertise it for server discovery
Problem:
Opening the new Server Dashboard screen in the iOS app (added in a recent app version) causes the app to crash immediately every time. This started right after I disabled Remote Access and blocked port 32400 externally — before that (port open + Remote Access enabled), the dashboard worked fine.
Other features (streaming, notifications, console/log viewer) all work correctly through the reverse proxy after enabling WebSocket forwarding on the proxy. Only the Dashboard screen crashes.
Troubleshooting done:
- Confirmed the crash is reproducible every time on iOS, both on home Wi-Fi (different network than the NAS, with no LAN access option in my setup) and mobile data
- Dashboard works fine when accessed via the web client (app.plex.tv or the reverse proxy domain directly in Safari)
- I suspect re-enabling Remote Access and opening port 32400 would make the dashboard work again, but I haven’t tested this directly for security reasons — this is an inference based on it having worked previously (when the port was open and Remote Access was enabled)
My guess is the Dashboard feature may be hardcoded to expect a reachable direct connection (Remote Access relay/direct path) rather than falling back to the custom server access URL / reverse proxy connection that works for everything else.
Would appreciate if someone from the team could take a look, or let me know if there’s a workaround short of re-opening direct server access.