How to Handle High Concurrent Load in Telegram Bots (1000+ Users Pressing Buttons Simultaneously)

[sanchik7777]

Hello,

I’m developing a Telegram bot using aiogram (version 3.20.0.post0). I'm also using the limited_aiogram library to throttle outgoing requests to the Telegram API according to official rate limits.

Right now, I’m working on optimizing the bot for high concurrent load. Imagine a situation where 1000+ users press an inline button at the same time. This button triggers a handler that does the following:

Fetches user data from the database

Calls callback.message.edit_text(...) to update the message

Sometimes calls callback.message.answer(...) to send a reply

Here’s my hypothesis:
If I understand the Telegram documentation correctly, both edit_text and answer count as outgoing messages and are subject to the 30 messages per second limit per bot (not per user).

So if 1000 users press the button at once, the bot will need approximately:

1000 / 30 = ~33 seconds

That means the last user might get a response only after 30+ seconds, which is unacceptable — users will start dropping off due to slow response.

Here are my questions:

Am I right in assuming that edit_text and answer count as outgoing messages under the 30 messages/sec bot-wide limit?

How do large bots (with 50,000–100,000+ users) handle this limitation?

Do they use queues, caching, aggregation strategies?

What are the best practices for cases when thousands of users are interacting at the same time, and each one needs personalized data?

I’d really appreciate any tips, examples, articles, or explanations on how large-scale Telegram bots manage to handle this kind of load while staying responsive.

Thanks in advance!

[Sarisan]

What are the best practices for cases when thousands of users are interacting at the same time, and each one needs personalized data?

Multithreading, asynchronous programming

[sanchik7777]

Thank you!
Am I correct in understanding that when dealing with a large number of concurrent users, it's better to use asynchronous processing (async/await) to efficiently utilize resources and avoid blocking operations, especially when working with the network or database?

And is multithreading useful for tasks that are CPU-bound or require parallel execution?

One more thing I'd like to clarify: is it true that methods like callback.message.answer() and callback.message.edit_text() are considered outgoing messages and fall under the 30 messages per second per bot limit imposed by the Telegram Bot API?
So if 1000 users press a button at the same time, the bot will only be able to respond at a rate of 30 messages per second, meaning the last user might have to wait over 30 seconds?

How do large bots handle this limitation — using queues, aggregation, caching, or even multiple bots?