bot stops responding after request timeout

[asserbrh]

the bot works, it is pretty intense in terms of requests, i made a throttler for it so we're not hitting rate limit, when we do in any case, we backoff and retry, still, as soon as I get this error:
"Request timeout. Request: method=post url=sendMessage params=<aiohttp.formdata.FormData object at 0x768818222150> files=None request_timeout=300"

I have tried stopping the polling and rerun it again, still no luck, as soon as a request times out once for sending a message, my bot freezes and stops sending or interacting.

[coder2020official]

does the polling continue

[All-The-Foxes]

does the polling continue

This also affects async. Polling continues, callback queries and message editing still works, but new messages will not send anymore and the bot has to be restarted. This might have something to do with why that's happening. Request timeouts are not being sent as the expected ApiTelegramException type.

ERROR [2025-12-30 18:59:37] Task exception was never retrieved
future: <Task finished name='Task-762' coro=<send_log() done, defined at /etc/service/hub/src/otterbot.py:247> exception=RequestTimeout('Request timeout. Request: method=post url=sendMessage params=<aiohttp.formdata.FormData object at 0x7f45285c9240> files=None request_timeout=300')>
Traceback (most recent call last):
File "/usr/local/lib/python3.14/dist-packages/telebot/async_telebot.py", line 3341, in send_message
await asyncio_helper.send_message(
...<5 lines>...
)
File "/usr/local/lib/python3.14/dist-packages/telebot/asyncio_helper.py", line 319, in send_message
return await _process_request(token, method_name, params=params, method='post')
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/dist-packages/telebot/asyncio_helper.py", line 105, in _process_request
raise RequestTimeout("Request timeout. Request: method={0} url={1} params={2} files={3} request_timeout={4}".format(method, url, params, files, request_timeout, current_try))
telebot.asyncio_helper.RequestTimeout: Request timeout. Request: method=post url=sendMessage params=<aiohttp.formdata.FormData object at 0x7f45285c9240> files=None request_timeout=300

I've set RequestTimeout to be explicitly caught everywhere I catch ApiTelegramExceptions and then wait a moment and retry. If that doesn't fix it I'll report back, otherwise no news is good news. But on the flip side, the only reason this can even be tested is because the API is acting goofy right now and dropping a lot of requests.

[SSxQ]

i also have the same issue
sometimes even every 3 message sent to bot fails with such error

no more than 10 seconds pass from the moment the message is sent until the error appears in the logs

maybe it requires some attempts or something like that?

i using pyTelegramBotAPI==4.29.1 and AsyncTeleBot

2026-01-07 20:49:09,827 (asyncio_helper.py:101 MainThread) ERROR - TeleBot: "Aiohttp ClientError: ClientOSError"

2026-01-07 20:49:09,827 (async_telebot.py:592 MainThread) ERROR - TeleBot: "Request timeout. Request: method=post url=sendMessage params=<aiohttp.formdata.FormData object at 0x00000280CA878D00> files=None request_timeout=300"

[Badiboy]

Most probably both of you have faulty code logic that stop event loops or processing loops or something else.

Could any of @asserbrh or @SSxQ provide full bot code that faces such issue? Without your private logic, just to have a reproduceable test case.

[All-The-Foxes]

In my case it's just a simple 'await bot.send_message()' call. I catch ApiTelegramExceptions, send them to my exception handler, then return a response on whether or not sending should be retried. Ever since I also set it to catch RequestTimeout in a secondary Except block this issue stopped happening. I think the reason is because RequestTimeout is its own separate exception that people aren't catching because we expect only ApiTelegramException, so the exception doesn't get caught and sending messages gets stuck.

[SSxQ]

I checked code in asyncio_helper.py and found this

current_try=0
session = await session_manager.get_session()
while not got_result and current_try<MAX_RETRIES-1:
    current_try +=1
    try:
        async with session.request(method=method, url=API_URL.format(token, url), data=params, timeout=timeout, proxy=proxy) as resp:
            got_result = True
            logger.debug("Request: method={0} url={1} params={2} files={3} request_timeout={4} current_try={5}".format(method, url, params, files, request_timeout, current_try).replace(token, token.split(':')[0] + ":{TOKEN}"))
            
            json_result = await _check_result(url, resp)
            if json_result:
                return json_result['result']
    except (ApiTelegramException,ApiInvalidJSONException, ApiHTTPException) as e:
        raise e
    except aiohttp.ClientError as e:
        logger.error('Aiohttp ClientError: {0}'.format(e.__class__.__name__))
    except Exception as e:
        logger.error(f'Unknown error: {e.__class__.__name__}')
    if not got_result:
        raise RequestTimeout("Request timeout. Request: method={0} url={1} params={2} files={3} request_timeout={4}".format(method, url, params, files, request_timeout, current_try))   

we have by default 3 attempts (MAX_RETRIES = 3) and INSIDE loop we do this

    if not got_result:
        raise RequestTimeout("Request timeout. Request: method={0} url={1} params={2} files={3} request_timeout={4}".format(method, url, params, files, request_timeout, current_try))

why? we dont do attempts at all in this case
also, why we do MAX_RETRIES-1 ?

[SSxQ]

for me, i did such fix (i dont know it is right or not, but now i dont have such problems with attempts, it works much better)

async def _process_request(token, url, method='get', params=None, files=None, **kwargs):
request_timeout = kwargs.pop('request_timeout', None)
request_timeout = REQUEST_TIMEOUT if request_timeout is None else request_timeout

params = _prepare_data(params, files)
timeout = aiohttp.ClientTimeout(total=request_timeout)

session = await session_manager.get_session()
last_exception = None

for current_try in range(MAX_RETRIES):
    try:
        async with session.request(
            method=method,
            url=API_URL.format(token, url),
            data=params,
            timeout=timeout,
            proxy=proxy
        ) as resp:
            logger.debug(f"Request attempt {current_try}: method={method} url={url}")
            json_result = await _check_result(url, resp)
            if json_result:
                return json_result['result']
    
    # do i need repeat such expections?
    except (ApiTelegramException, ApiInvalidJSONException, ApiHTTPException) as e:
        raise e

    except aiohttp.ClientError as e:
        last_exception = e
        logger.warning(f"Aiohttp ClientError on attempt {current_try}: {e}")

    except Exception as e:
        last_exception = e
        logger.warning(f"Unknown error on attempt {current_try}: {e}")

    # TODO: rewatch, if i really need sleep here
    if current_try < MAX_RETRIES:
        await asyncio.sleep(2 ** current_try)

raise RequestTimeout(
    f"Request timeout after {MAX_RETRIES} attempts. "
    f"method={method} url={url} params={params} files={files}. "
    f"Last exception: {last_exception}"
)

[SSxQ]

Hi, i also checked synchronous version of telebot and it have the right way of checking result OUTSIDE while loop
Here is the code

elif RETRY_ON_ERROR and RETRY_ENGINE == 1:
        got_result = False
        current_try = 0
        while not got_result and current_try<MAX_RETRIES-1:
            current_try+=1
            try:
                result = _get_req_session().request(
                    method, request_url, params=params, files=files,
                    timeout=(connect_timeout, read_timeout), proxies=proxy)
                got_result = True
            except HTTPError:
                logger.debug("HTTP Error on {0} method (Try #{1})".format(method_name, current_try))
                time.sleep(RETRY_TIMEOUT)
            except ConnectionError:
                logger.debug("Connection Error on {0} method (Try #{1})".format(method_name, current_try))
                time.sleep(RETRY_TIMEOUT)
            except Timeout:
                logger.debug("Timeout Error on {0} method (Try #{1})".format(method_name, current_try))
                time.sleep(RETRY_TIMEOUT)
        if not got_result:
            result = _get_req_session().request(
                    method, request_url, params=params, files=files,
                    timeout=(connect_timeout, read_timeout), proxies=proxy)

I also wonder, here we have MAX_RETRIES = 15, and in async version only 2 attempts (if it worked)
Why are there so many fewer attempts?

[coder2020official]

Hey, thanks for the update, I will reinvestigate this issue soon