Error Handling

Errors can be correctly handled with try...except blocks in order to control the behaviour of your application. Hydrogram errors all live inside the errors package:

from hydrogram import errors

RPCError

The father of all errors is named RPCError and is able to catch all Telegram API related errors. This error is raised every time a method call against Telegram’s API was unsuccessful.

from hydrogram.errors import RPCError

Warning

Avoid catching this error everywhere, especially when no feedback is given (i.e. by logging/printing the full error traceback), because it makes it impossible to understand what went wrong.

Error Categories

The RPCError packs together all the possible errors Telegram could raise, but to make things tidier, Hydrogram provides categories of errors, which are named after the common HTTP errors and are subclass-ed from the RPCError:

from hydrogram.errors import BadRequest, Forbidden, ...

Single Errors

For a fine-grained control over every single error, Hydrogram does also expose errors that deal each with a specific issue. For example:

from hydrogram.errors import FloodWait

These errors subclass directly from the category of errors they belong to, which in turn subclass from the father RPCError, thus building a class of error hierarchy such as this:

  • RPCError
    • BadRequest
      • MessageEmpty

      • UsernameOccupied

      • ...

    • InternalServerError
      • RpcCallFail

      • InterDcCallError

      • ...

    • ...

Unknown Errors

In case Hydrogram does not know anything about a specific error yet, it raises a generic error from its known category, for example, an unknown error with error code 400, will be raised as a BadRequest. This way you can catch the whole category of errors and be sure to also handle these unknown errors.

Errors with Values

Exception objects may also contain some informative values. For example, FloodWait holds the amount of seconds you have to wait before you can try again, some other errors contain the DC number on which the request must be repeated on. The value is stored in the value attribute of the exception object:

import asyncio
from hydrogram.errors import FloodWait

...
    try:
        ...  # Your code
    except FloodWait as e:
        await asyncio.sleep(e.value)  # Wait N seconds before continuing
...