telegram failing

[hollow creek]

"Telegram plugin fails with ETIMEDOUT in Docker host-network mode on Unraid. Node.js fetch to api.telegram.org works fine but the grammy polling fails. OpenClaw version 2026.4.24."

[sick pasture]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[crisp vector]

This smells like the getUpdates long-poll path, not basic Telegram reachability. A simple Node fetch("https://api.telegram.org")/getMe can pass while grammY/OpenClaw polling still times out, because polling holds a long HTTPS request and uses OpenClaw’s Telegram transport/undici dispatcher.

I’d try this order:

openclaw channels status --probe
openclaw channels logs --channel telegram
openclaw logs --follow

Then in Docker/Unraid, force the Telegram path toward IPv4 and restart the container:

{
  channels: {
    telegram: {

      network: {
        autoSelectFamily: false,
        dnsResultOrder: "ipv4first",
      },
    },
  },
}

or as env:

OPENCLAW_TELEGRAM_DISABLE_AUTO_SELECT_FAMILY=1
OPENCLAW_TELEGRAM_DNS_RESULT_ORDER=ipv4first

Also useful: enable targeted Telegram diagnostics:

{
  diagnostics: {
    flags: ["telegram.http"]
  }
}

Docs:

• https://docs.openclaw.ai/channels/telegram
• https://docs.openclaw.ai/channels/troubleshooting
• https://docs.openclaw.ai/diagnostics/flags

If it still fails, test the actual long-poll from inside the same container/network namespace, not just a generic fetch/getMe. Don’t paste the token here, but run something like: