openclaw - 💡(How to fix) Fix Telegram polling provider startup loops when deleteWebhook cleanup fails [1 comments, 2 participants]

Summary

On Windows with OpenClaw 2026.4.26 (be8c246), enabling the Telegram channel can make the gateway intermittently unresponsive during startup because the Telegram polling runtime treats deleteWebhook cleanup failure as a setup retry loop.

This happens even when no Telegram webhook URL is configured and the intended mode is polling.

Environment

• OpenClaw: 2026.4.26 (be8c246)
• OS: Windows 10.0.26200
• Node runtime in logs: 24.14.1
• Gateway bind: local loopback, port 18790
• Telegram channel: channels.telegram.enabled=true
• Telegram bot token/API basics were validated separately with Telegram API getMe and getWebhookInfo returning HTTP 200. Webhook URL was empty and pending updates were 0.

Observed behavior

When Telegram is enabled, gateway startup reaches Telegram provider startup and then logs errors like:

telegram deleteWebhook failed: Network request for 'deleteWebhook' failed!
Telegram webhook cleanup failed: Network request for 'deleteWebhook' failed!; retrying in 3.69s.

During this retry path, the gateway can become hard to probe from the CLI:

Error: gateway timeout after 60000ms
Gateway target: ws://127.0.0.1:18790
Source: local loopback
Bind: loopback

Keeping Telegram disabled and/or skipping channel startup restores gateway stability:

Gateway Health
OK (~300ms)
Local loopback ws://127.0.0.1:18790
  Connect: ok (~25-40ms) · Capability: admin-capable · Read probe: ok

Code path observed in the installed package

In the installed package, Telegram polling cleanup appears to unconditionally call deleteWebhook before polling can proceed:

async #ensureWebhookCleanup(bot) {
  if (this.#webhookCleared) return "ready";
  try {
    await withTelegramApiErrorLogging({
      operation: "deleteWebhook",
      runtime: this.opts.runtime,
      fn: () => bot.api.deleteWebhook({ drop_pending_updates: false })
    });
    this.#webhookCleared = true;
    return "ready";
  } catch (err) {
    return await this.#waitBeforeRetryOnRecoverableSetupError(err, "Telegram webhook cleanup failed") ? "retry" : "exit";
  }
}

Installed dist path:

dist/extensions/telegram/monitor-polling.runtime-*.js

I could not find a documented config/environment option to skip this cleanup step.

Expected behavior

For polling mode, deleteWebhook cleanup should not make the whole Telegram provider/gateway effectively unusable when it fails transiently.

Possible fixes:

1. Add a documented option such as channels.telegram.skipWebhookCleanup or OPENCLAW_TELEGRAM_SKIP_WEBHOOK_CLEANUP=1.
2. Treat deleteWebhook cleanup failure as non-fatal in polling mode after logging a warning, then continue to getUpdates.
3. Only run cleanup when a webhook was actually configured or when getWebhookInfo shows an active webhook.

Current workaround

Keep Telegram disabled or skip channel startup. This keeps the gateway stable but makes Telegram unavailable.

Security note

No bot token, gateway token, or sender IDs are included here.