Make logging API/SDK timeouts uniform before release

[carlosalberto]

Before going stable, there's a chance to review and align the timeout values:

• In traces, export and flush timeouts default to 30 seconds, whereas in metrics they default to 10 seconds. In logging most places use 10, but one-two use 30 seconds instead. They should all get the same value.
• In some methods, timeout can receive a None value where in other places it can only get an int. Whatever decision is taken, all methods should align to that:

def force_flush(self, timeout_millis: int | None = None):
# VS
def force_flush(self, timeout_millis:int = None):

[saisharan0103]

The repro details point to two consistency gaps: timeout defaults differ across traces, metrics, and logging, and force_flush-style signatures do not agree on whether None is part of the public type. For a PR, I would start by choosing one logging timeout default and then updating the annotations/defaults together so runtime behavior and typing match. The tests should cover both the default value and explicit None handling.