Telegram error 400: Bad Request: bad webhook: Webhook can be set up only on ports 80, 88, 443 or 8443

The URL given to setWebhook includes an explicit port that is not one of the four Telegram supports. The Bot API documentation and FAQ both list them: 443, 80, 88 and 8443. Any other port, most commonly 3000, 5000, 8000 or 8080 from a development server, is rejected at registration time.

The description includes the literal port list, so you will sometimes see it phrased in guides as port 8080 is not supported; the actual text returned is the one above. Only a local Bot API server lifts this restriction ('use any port for the webhook').

Because the webhook was never registered, Telegram keeps delivering to the previous webhook or to nothing. getWebhookInfo will show the old state, which is the fastest way to confirm what actually happened after a failed registration. If your bot previously used polling and this setWebhook failed, the bot silently keeps polling mode, and a later poller may hit the 409 webhook conflict once someone registers the corrected URL.

The error appears in libraries as a plain 400: python-telegram-bot raises telegram.error.BadRequest from bot.set_webhook(...), aiogram raises TelegramBadRequest, and the Node libraries reject the setWebhook promise with the description text. Frameworks that auto-register webhooks on boot often log and swallow this, which turns a one-line configuration mistake into a bot that starts 'successfully' and never receives an update. Make webhook registration a hard-fail step in deployment.

To reproduce and confirm the rule, call setWebhook with an explicit port and read the body: curl "https://api.telegram.org/bot<TOKEN>/setWebhook?url=https://example.com:8080/hook" returns exactly this description, while the same URL on 8443 registers and shows up in getWebhookInfo.

{"ok":false,"error_code":400,"description":"Bad Request: bad webhook: Webhook can be set up only on ports 80, 88, 443 or 8443"}

Separate the internal listening port from the public webhook URL in configuration, and validate the public URL's port against the allowed list at startup. Document the four ports in your runbook so the next deploy does not repeat the mistake. The full webhook constraint table (scheme, ports, max_connections, secret_token) is under webhook limits, and endpoint-side failures are covered in the webhook debugging guide.