sock.connect hang

I encountered a problem when using asyncsender, the fastapi project, fluent-logger==0.9.6

When I was running pytest, I found that it would be blocked for a long time after the pytest ended

And not any log to tell me why, stuck here for a while time

Found that when the logger sends logs, each log will be blocked for 20 seconds because of socket connect.

Because the fluent address error, every time when `_send_loop` consume a log, it will trigger `_reconnect`.

And each `_reconnect` will took 20 seconds. I thought of two ways
- maybe there is a way that we can add timeout on connect
- when we meeted this kind error, we clear the queue and print some log

Sorry about the bad english :). Feel free to close this if un-needed :)



Start by tracing the logger's _send_loop and _reconnect paths, then reproduce the behavior with pytest and an invalid Fluent address. Check how socket connection failures block log consumption and verify the chosen handling prevents repeated 20-second stalls while providing the reported diagnostic behavior.

Written by the indexing model from the issue text.