telegraf: Should handle 429 flood wait errors during polling

Context

Telegraf.js Version: 4.4.2
Node.js Version: 17.1.0
Operating System: alpine

Minimal Example Code Reproducing the Issue

const { Telegraf } = require('telegraf');

const bot = new Telegraf(process.env.TELEGRAM_BOT_TOKEN);

bot.start((ctx) => ctx.reply('Welcome'));

bot.on('text', async (ctx) => { ...

Not sure how to reproduce the issue, but here is a stack trace

2021-12-30T15:32:31.446352795Z /usr/src/app/node_modules/telegraf/lib/core/network/client.js:263
2021-12-30T15:32:31.446436840Z             throw new error_1.default(data, { method, payload });
2021-12-30T15:32:31.446457232Z                   ^
2021-12-30T15:32:31.446472689Z
2021-12-30T15:32:31.446487615Z TelegramError: 429: Too Many Requests: retry after 5
2021-12-30T15:32:31.446502487Z     at Telegram.callApi (/usr/src/app/node_modules/telegraf/lib/core/network/client.js:263:19)
2021-12-30T15:32:31.446517557Z     at runMicrotasks (<anonymous>)
2021-12-30T15:32:31.446532914Z     at processTicksAndRejections (node:internal/process/task_queues:96:5)
2021-12-30T15:32:31.446548317Z     at async Polling.[Symbol.asyncIterator] (/usr/src/app/node_modules/telegraf/lib/core/network/polling.js:27:33)
2021-12-30T15:32:31.446563300Z     at async Polling.loop (/usr/src/app/node_modules/telegraf/lib/core/network/polling.js:69:30) {
2021-12-30T15:32:31.446578644Z   response: {
2021-12-30T15:32:31.446592851Z     ok: false,
2021-12-30T15:32:31.446607299Z     error_code: 429,
2021-12-30T15:32:31.446621857Z     description: 'Too Many Requests: retry after 5',
2021-12-30T15:32:31.446667056Z     parameters: { retry_after: 5 }
2021-12-30T15:32:31.446685253Z   },
2021-12-30T15:32:31.446699808Z   on: {
2021-12-30T15:32:31.446714119Z     method: 'getUpdates',
2021-12-30T15:32:31.446728450Z     payload: { timeout: 50, offset: 462581445, allowed_updates: [] }
2021-12-30T15:32:31.446743346Z   }
2021-12-30T15:32:31.446757413Z }

Expected Behavior

I assume that polling should swallow this 429 and silently continue working after this response from Telegram

About this issue

Original URL
State: closed
Created 3 years ago
Comments: 16

Commits related to this issue

Retry `getUpdates` upon 429; fixes #1563 — committed to PabloWenger/telegraf by wojpawlik 2 years ago

Most upvoted comments

Apparently the method getUpdates is only specified to never return 429 errors during normal operation of the Bot API server. However, it can indeed return 429 when the Bot API server is closing.

Consequently, telegraf should be changed to handle flood wait errors during long polling by checking the error status code for containing 429, and delaying subsequent calls to getUpdates by the specified number of seconds.

KnorpelSenf on Jan 22, 2022

This error is explicitly specified not to happen. Telegraf already handles all possible errors correctly. It doesn’t make sense to deviate from their specification. I forwarded this issue to Telegram.

It’s not a good first issue to work on for any new contributors because this cannot be solved easily. In fact, this might not be possible at all to be solved on our end. It requires some level of understanding about how the Bot API server works, and how it communicates with Telegram via MTProto.

KnorpelSenf on Jan 22, 2022

Check this I recommend you use https://npm.im/telegraf-throttler

hypernova7 on Jan 3, 2022