Hello again,

Concerning API shutdowns, it seems that changing request intervals to 600ms did the trick, and keep API alive and working all night long.

Maybe if it's not too complicated, you can add an adaptive "extra delay" when api errors persist more than a few seconds.
It seems to me that client keep trying to connect and BTC persists rejecting longer each time, so when shutdown time is large (9999...) you must wait some 5 - 10 minutes before being allowed to connect again.

I can perfectly live with the program as it works now, but maybe it could be helpful on high traffic scenarios.

Again thanks for your excelent help and support.
Best whishes
Ricardo