grim elm
Node 24.14.1 on Windows 11 , Telegram getUpdates long-poll stalls every ~180s with Network request failed. sendMessage also fails intermittently. curl works fine. Happens on both WiFi and mobile hotspot. OpenClaw 2026.4.25.
What happened:
Everything was working fine until I installed Claude Code Premium. Since then, Telegram polling stalls and sendMessage fails intermittently.
Symptoms:
getUpdateslong-poll stalls every ~180s then crashes withNetwork request for 'getUpdates' failed!sendMessagefails withNetwork request for 'sendMessage' failed!- Gateway auto-restarts the polling loop but it stalls again every time
- Bot receives messages occasionally but can never send replies back
- Group messages show
skipping group message — reason: not-allowed(expected, DM policy)
What I've ruled out:
- ✅ curl to api.telegram.org works fine (got 302)
- ✅ DNS resolves correctly (Videotron ISP, Montreal)
- ✅ No VPN active
- ✅ Windows Defender only, no third-party AV
- ✅ Tested on both WiFi AND mobile hotspot — same result
- ✅ No Docker, no WSL2 — OpenClaw runs native Node on Windows
- ✅ Config is valid, health shows OK (green)
- ✅ IPv4 first forced via NODE_OPTIONS — no change
Key log lines:
[telegram] Polling stall detected (active getUpdates stuck for 183s); forcing restart.
[telegram][diag] polling cycle finished reason=polling stall detected error=Network request for 'getUpdates' failed!
telegram sendMessage failed: Network request for 'sendMessage' failed!
failed to persist update offset: Error: EBUSY: resource busy or locked
Question:
Is there a known fix or workaround for long-poll TLS connection drops on Node 24 + Windows? Any config flag or env variable that can help stabilize the connection?
Thanks 🦞
elder shoal