LogtailHandler: implement logging.Handler.flush

[PetrHeinz]

Implementing flush method on our handler.

The method should be implemented by subclasses of logging.Handler. It's called automatically during graceful termination.
https://github.com/python/cpython/blob/8e8a4baf652f6e1cee7acde9d78c4b6154539748/Lib/logging/__init__.py#L1039-L1046

The method will enforce uploading the current buffer to Better Stack sooner than in flush_interval, and will wait until all data are sent to Better Stack.

---

Making check_interval configurable and decreasing the default value from 1s to 100ms, as it may have serious impact on short-running code (example project took ~1.3s to finish before this PR, now takes ~350ms).

---

It is now safe to terminate execution immediately (eg. using os._exit), if flush() is called before:

for handler in logger.handlers:
    handler.flush()

os._exit(os.EX_OK)

[PetrHeinz]

Seems that the 3.11 tests fail on master as well: https://github.com/logtail/logtail-python/actions/runs/9581314199/job/26417818292

[curusarn]

Looks good, thank you!